Bernard Malamud Letter Network
Philip Roth's Correspondence with Malamud
Geographic Map of Malamud's Letter Network
Flannery O'Connor's Correspondence with Malamud
Directed Graph of Malamud's Letter Network
Within this project you will find a collection of letters to and from Bernard Malamud. It is our hope that one day this project will encompass all of Malamud's correspondence. For now, we are limited in scope to the collections at the Library of Congress and the Harry Ransom Center at UT-Austin. However, we do not have all of Malamud's letters at UT-Austin. Most of what we currently have access to are letters between Malamud and his publishers, fans, and some public figures. Though Malamud wrote his brother Eugene more often than anyone else, none of their conversations appear in these collections because we were unable to obtain these letters from the UT-Austin Harry Ransom Center. In the network diagram, we illustrate this by including disconnected nodes for correspondents that we know Malamud interacted with, but that we were unsure of the frequency of letters exchanged. For this reason and others like it, our project does not claim to present a comprehensive picture for all of Malamud's correspondence. As Miriam Posner suggests, we must recognize "simultaneously the value of any particular dataset and its inevitable poverty, compared with the phenomena it purports to describe."
We organize these letters using three different interfaces: a collection of the letters themselves and two "distant reading" visualizations with which users can interact, a geographic map and a directed network graph.
- The collection of the letters is split out by correspondent, and includes short biographies of select public individuals. Future work will include tracking down biographical information for as many correspondents as possible. If you are looking to view individual letters, head here. We hope to eventually include transcriptions of letters as well, but currently there are only photos of the physical letters. We made this decision to focus on collecting a large number of letters rather than including a smaller number of transcribed letters both because of time constraints and a desire to implement the two following visualizations, which require many letters to be effective.
- The geographic map pinpoints the locations from which letter was sent, and helps to show how Malamud interacted with different individuals that tended to cluster in specific locations. Zoom in on pinned locations to see more detailed letter distribution information, including multiple pins spiraling from a single location. Each pin links to the letter page, so feel free to click on a pin to move between the visualization and the letter photos. Only letters that include spatial metadata are mapped. There are two versions of the map included.
- The first version is built natively in Scalar using the Google Maps layout, but our group ultimately deemed this visualization unsatisfactory because it does not adequately represent marker density or overlapping markers; high density areas were not immediately distinguishable from low density areas when zoomed out, and overlapping markers or markers with the exact same set of lat/long coordinates were depicted as a single marker. We are including this version to show how the second version improves upon it.
- The second version is built using Google Maps' Javascript API, Scalar's API, and the MarkerClusterer (MC) and OverlappingMarkerSpiderfier (OMS) libraries; it is hosted externally on CodePen. This map achieves the same easily identifiable Google Maps layout as the native Scalar version, but improves upon it by implementing the MC and OMS libraries. The MC library clusters together markers at a low zoom scale so that marker density - in this case, representing letter frequency - is easily identifiable. Once a user zooms in, the clusters revert to individual markers to allow for individual interaction. Overlapping markers or markers with the same lat/long coordinates are helped by the OMS library, which "spiderfies" overlapping markers when users click on them so that all overlapping markers are displayed simultaneously, facilitating a more accurate representation of letters.
- The directed graph is built using Scalar's API and the SigmaJS Javascript graphing library; it is hosted externally on CodePen. This graph shows how Malamud interacted with individual correspondents by displaying each correspondent as a node and letters between correspondents as edges – lines between the nodes. The number of letters a correspondent sent and received influences the each node’s size and the size of both sending and receiving edges, respectively. The forceAtlas algorithm used to shape the graph pushes nodes with stronger relationships closer together and disconnected nodes further apart. This graph ultimately uses Malamud’s correspondence as a proxy for relationship strength. In the future, we hope to add node colors to distinguish relationship types such as “fan,” “literary figure,” “public figure,” “university administration,” “publisher,” “family,” and “friend.” Right now, grey nodes indicate “disconnected” nodes for correspondents we know existed but do not have letter images for. Click and drag a node to see it better; it will drag related nodes with it. Double click a node to go to the appropriate Scalar page for that correspondent or the SCARC or UT-Austin archive page. Edge weights are labeled, and edge direction for whether the edge represents letters sent or received is indicated by arrows on the end of edges (zoom closely if you can’t see them).
As a whole, we hope our project provides multiple perspectives for reading and analyzing Malamud's correspondence, giving users a better picture and understanding of Malamud's life, his publishing process, the fans and public figures who wrote him, and the letter writing practices common during Malamud's life.