Telegraph code books, in the hundred years of their use (ca. 1850-1950), served primarily as an information compression technology, matching a phrase or sentence with a code word, thus saving money on telegram costs. On occasion, they were also used for secrecy, since typically only someone possessing the same code book as the sender would be able to decode the message. In the books’ heyday, virtually every industry had its specialized version; shipping, banking, railways, carpets, and rubber all had code books designed especially for them, as did many other industries as well. In addition, general purpose code books also flourished, many going through multiple editions. The code books were widely used, especially by businesses but also private individuals. Estimates are that 95% of all transatlantic telegraph traffic was encoded, and a high percentage of domestic traffic as well. After the telegraph became obsolete, most libraries and individuals purged the books from their collections; of the hundreds of thousands of originals, perhaps only hundreds remain. Old enough to have progressed from junk to collectible, they have now become prized items and sell for fancy prices at online auction sites.
Chapter 5 of How We Think traces the evolution of telegraph codes from their human-centric beginnings to the increasingly machine-centric codes constructed algorithmically in the late nineteenth and early twentieth centuries. Their significance lies partly in this genealogy; through the Baudot code and the Teletype machine, they became the ancestors of the ASCII code still used in digital computers today. In addition to their implications for media theory, they have potent historical significance, for they preserve, like a fly in amber, messages that the compilers thought would be useful in a very wide variety of situations, thus opening a window onto how people thought and wrote over the century of their use. Until now, however, this material has remained virtually untouched with the notable exceptions of a few media historians such as David Kahn and John McVey, because the digitized scans could only be accessed individually, with no easy way to go from book to book or to compare groups of books through machine analysis. Historians, cultural critics and other potential users in general do not consult them, because to do so would require going through book after book by hand, for what may in the end be only a paragraph or footnote in their research projects.
This section of How We Think: A Digital Companion aims to remedy the situation by providing an accessible, easy-to-use interface that enables users to search for words through the entire collection of scans with a single keystroke. Entering a word brings up a list of all the code books where that word appears, and one can then navigate via the scan at our site or click to be taken to where the book scan is stored at Google Books or the Internet Archive. Moreover, this interface makes possible comparative word frequency analysis on all the available scans, thus enabling interested parties to ascertain which words or phrases where most widely used in which books. For convenience, a list of the top 300 words, as determined by weighted frequencies, is provided for the user’s information, as well as a list of intriguing bigrams (two words found frequently together). Other kinds of machine analytics are of course also possible.
Through the generosity of Nicholas Gessler and the Perkins Library at Duke University, this project has also added to the archive twenty-two full-text scans of code books published before 1923 that are in the Gessler Collection, a contribution made more valuable because these are not excerpts or selections, as is the case for many telegraph code books in Google Books and the Internet Archive, but full-text scans. As I discovered as a result of this project, scanning one of these books is far from trivial. Many of the pages have bleed-through from the other side, making a clear scan difficult, and for the older books, the pages are often thin, fragile, and tricky to handle. I am grateful to Deborah Jakubs, the Rita DiGiallonardo Holloway University Librarian and Vice Provost for Literary Affairs, and Winston Atkins at the Duke University Library for their help in arranging and paying for the scans to be made, and to Nicholas Gessler for allowing his books to be included.
Many of the code books are available as “plain text” for further analysis. A list of distinctive words is also available. Examples of distinctive words include: