This is the second of a two-part series on free digital archives featuring English ballads. It follows Eleanor’s discussion of the JISC-funded Broadside Ballad Initiative at Oxford.
The University of California at Santa Barbara has created a free digital ballad collection called The English Broadside Ballad Archive (EBBA), which provides access to more than 8,000 seventeenth-century ballads. The collection includes ballads from the Pepys Collection, the Roxburgh Collection, the Euing Collection, and the Huntington Library. EBBA is directed by Patricia Fumerton at UCSB. This project was supported by the N.E.H.
Individual entries provide links to sheet facsimiles, facsimile transcriptions, and often recordings. These features facilitate introducing students both to ballads’ visual details–ornaments, woodcuts, columned verse–and to their tunes.
Cataloging is full and includes the following:
EBBA ID: An internal identifier. Each individual ballad in the archive has a unique EBBA ID.
Title: A diplomatic transcription of the ballad title as it appears on the ballad sheet. The title consists of all ballad text before the first lines of the ballad, including verse headers but excluding text recorded elsewhere under other catalogue headings (such as the license or author, date, publisher and printer imprints).
Date Published: The year—or, in most cases, range of years—during which EBBA believes the ballad to have been published. See Dates.
Author: The recognized author of the ballad in cases where an indication of authorship has been printed on the ballad or, in the case of Pepys ballads, when Weinstein has identified an author from external sources (e.g., Wing, Rollins).
Standard Tune: The standardized name for the melody (according to Claude M. Simpson or other reliable sources). Clicking the standard tune name will return all ballads with the same melody, including alternate tune titles.
Imprint: A diplomatic transcription of the printing, publishing, and/or location information as it appears on the ballad sheet.
License: A diplomatic transcription of the licensing or permission information as printed on the ballad.
Collection: The name of the collection to which the ballad belongs. In cases where the ballad is not part of a named collection, the name of the holding library plus “miscellaneous” will appear. For example, Huntington Library ballads that are not part of a collection are grouped as “HEH Miscellaneous.”
Sheet/Page: For ballads that are collected as independent sheets, the citation page displays the word “Sheet” and lists the sheet number given to it by its holding institution (usually part of its shelfmark). For ballads bound in a book, the citation page displays the word “Page” and lists the page number within the bound volume.
Location: The name of the holding institution.
Shelfmark: The shelfmark assigned by the holding institution.
ESTC ID: The Citation Number for the English Short Title Catalogue (ESTC). Use this number to find the full ESTC citation for any given ballad at http://estc.bl.uk/.
Keyword Categories: The keywords from EBBA’s standardized keyword list that relate to the ballad’s theme and content.
Notes: Clarify potential areas of confusion for users, such as ballads that have print on both sides of a sheet.
MARC Record: A link to our MARC-XML records
Additional Information: Information specific to each part of the ballad.
Title: Separate titles for multi-part ballads.
Tune Imprint: Tune title(s) as printed.
First Lines: A diplomatic transcription of the first two lines of the ballad text proper, below any heading information included in the title or elsewhere under other catalogue headings.
Refrain: Repeated lines at the end of or within ballad stanzas.
Condition: Description of ballad sheet damage and the current state of the sheet. (This information is from Weinstein and is currently for the Pepys collection only.)
Ornament: A list of decorations made of cast metal that appear on the ballad. Frequently used to fill empty spaces in the forme and/or to delimit parts of the ballad text, these ornaments include vertical rules, horizontal rules, and cast fleurons. (This information is from Weinstein and is currently for the Pepys collection only.)
Ballad scholars working with EEBO or ECCO will be familiar with the difficulty of finding ballads, making English Broadside Ballad Archive and Bodleian Library Broadside Ballads necessary.
Together with new printed resources, such as Patricia Fumerton and Anita Guerrini’s Ballads and Broadsides in Britain, 1500-1800 (Ashgate 2010) and Angela McShane’s Political Broadside Ballads of Seventeenth-Century England: A Critical Bibliography (Pickering & Chatto 2011), these digital resources provide a robust and growing archive for the systematic study of a format whose transiency may have discouraged such studies in the past.
Early Modern Online Bibliography 
February 25, 2013 at 1:25 pm |
Many thanks, Anna, for drawing specific attention to EEBA and offering these extremely informative details about this resource. The growing critical work on ballads and their increasing availability through projects such as EEBA speak to the long overdue recognition that ballads and other forms of ephemera are receiving. Book History has long been more about just “books,” and projects such as these give witness to the broader, inclusive nature of the field
As this post indicates, the study of ballads has steadily progressed since the publication of Diane Dugaw’s Warrior Women and Popular Balladry, 1650-1850 (University of Chicago Press, 1989).
LikeLike
February 25, 2013 at 7:59 pm |
Thank you for calling attention to EBBA’s deep and multi-media resources. The EBBA team is currently archiving the approximately 1,700 broadside ballads (published mostly pre-1700) in the Crawford and other collections at the National Library of Scotland, made possible with another NEH grant. Cataloging of the NLS collections is now complete and images will be up soon. We have also received an NEH start-up grant for a highly effective image searching tool–the brain-child of Carl Stahmer–and more extensive cataloging of the ballad woodcut illustrations–the brain-child of Megan Palmer Browne. I and ten others of the EBBA team have furthermore just brought out a companion to the over 1,800 broadside ballads collected by Samuel Pepys and archived in EBBA.The edition includes introductory essays on ballad culture and on Pepys’s collecting categories as well as 155 facsimile transcriptions and two CD recordings of all extant tunes for the selected ballads. The book is available at http://acmrs.org/publications/catalog/broadside-ballads-pepys.
LikeLike
February 25, 2013 at 9:21 pm |
Thanks for this. I can see how the new book serves as a guide to the larger database. These are all promising resources.
Can you tell us a bit more about how the image-searching tool works and how it compares to the Bodleian Library Broadside Ballad’s iconographic indexing?
LikeLike
February 25, 2013 at 9:23 pm |
Many thanks, Patricia, for these additional details. That you have published a companion text, Broadside Ballads from the Pepys Collection: A Selection of Texts, Approaches, and Recordings, Ed. Patricia Fumerton. MRTS Texts for Teaching, Vol. 6 (ACMRS Punblications, 2012, shows foresight in helping others see how this resource can be used most effectively as a pedagogical as well as a research tool.
It is nice to see Carl Stahmer’s name as someone involved; his computer know-how was quite important in launching Romantic Circles years ago. What is the relationship—if any—between Carl’s image-searching tools and that being used for the Oxford Ballad Collections?
I would also like to hear more about Megan Palmer Brown’s cataloguing of the illustrations in terms of subject matter, parameters for selecting keywords, and the like.
LikeLike
February 28, 2013 at 12:37 am |
Thanks much for your EBBA write-up!
Cataloguing these illustrations has been quite a journey—because of their ephemeral nature, their astonishing number, and the far-flung locations in which they’re collected, they haven’t been much studied (Samuel Pepys himself tried to classify them, but gave up). I came to UC Santa Barbara in 2006, and Patricia and Carl were already then discussing how a combination of traditional cataloguing and image-matching software could be harnessed to give these vibrant and quirky illustrations their due. So, our cataloguing and computation practices have been some years in the making!
On the cataloguing side: We experimented extensively with a number of models, including full-scale verbal description of the images and various keywording schemes. But we’ve come to the conclusion that the most useful, neutral, and interoperable model is one which will do two things: first, tag salient features of an image (key figures, features, and objects, drawn from a controlled vocabulary list created by head cataloguers in consultation with art historical standards); and, second, indicate in general terms the genre of the image (drawn from a Getty controlled vocabulary list—”portrait,” “narrative,” etc.).
Many illustrations appear numerous times across ballad archives, and for most, both “tag” and “genre” terms will be the same across woodcut impressions, even when a particular illustration appears in differing contexts (for example: an impression of a woman with a feathered fan appears both as a romantic heroine and as the notorious Wife of Bath). To acknowledge such differences, we’ll also give users access to the keywords associated with each ballad on which a particular impression appears, so that continuities or changes in the topical meaning of an image can be mapped.
LikeLike
February 28, 2013 at 8:56 am |
Many thanks, Megan, for these details about the cataloguing. Did you/your team design the first controlled list working with head cataloguers, etc. or did your team employ an existing list?
LikeLike
March 1, 2013 at 4:04 am |
You’re most welcome, Eleanor! Yes, we designed our first list, which we’re calling “descriptive tags,” ourselves: many of these terms came out of the ultimately-cumbersome verbal description experiment—certain key words appeared again and again; we culled these into a list and, after much spirited discussion (and consultation with other controlled vocabularies and art historical resources), we pared them down into a managable set which is also keyed to the specific character of our archive. So, for example, we have tags to identify some basic features of early modern dress with which non-expert viewers might not be familiar.
LikeLike
March 1, 2013 at 7:45 am |
The system for searching woodcuts seems promising, not least because the controlled vocabulary is not overly precise, which forces readers to spend time with the ballads. A search for woodcuts using the keyword “royalty” calls up 139 ballads in the Pepys collection. Sifting through those ballads is in itself an education.
LikeLike
March 3, 2013 at 7:47 pm |
Thanks, Megan. I’ve just begun trying to construct keywords to identify 320 or so late 18th-century illustrations, so I am very interested in your process/experience. Projects such as yours signal some exciting prospects for work in image-text relations.
LikeLike
March 3, 2013 at 7:57 pm |
Hi again, Anna and Eleanor. I should note, the keywords currently available on our site are an earlier iteration; we hope to roll out the new system late this summer. And thanks—we’re excited, too!
LikeLike
March 3, 2013 at 9:07 pm
Thanks for the clarification–that’s good to know.
LikeLike
March 25, 2013 at 7:04 pm |
In response to both Anna and Eleanor’s specific query about the how EBBA’s image-searching tool works and how it compares to the system being developed at the Bodleian Library, the short answer is that it is very similar. We have been in close contact with the folks at the Bodleian Library throughout the development process, and the two systems aim to provide similar functionality within their own archival universes as well as, ultimately, the ability to search across archives. With regard specifically to the EBBA system, which we call Archive-Vision, or Arch-V, users will be able use an existing image (or selected sub-region of an image) as a search seed, and the system will automatically search the entire archive for all occurrences of the seed in the archive. This will allow users to both find every occurrence of a particular representation in the archive and also to find, with a high level of accuracy, all impressions printed using the same physical block. Pending future funding, we hope to expand upon the latter of these functionalities by automatically matching the impressions with surviving, physical blocks.
Visual searching has obvious value for any scholar interested in the images that appears on the broadsides. But at EBBA we see this functionality not as a replacement for traditional cataloguing but rather as a potential enhancement to it. As such, we are using the tool as an assistive technology as part of our human cataloguing practice. Using the tool, cataloguers can see all occurrences of similar impressions when they are cataloguing a particular instance, including the ability to view, compare, and import the various keyword cataloguing (described by Megan above) for each similar impression. This process will dramatically improve the consistency of cataloguing across similar items in the archive.
On the technical side, Arch-V is a c++ application that also has java and PhP components that allow easy interface with web applications. It works by constructing a Visual Dictionary of extracted shapes (contours) that appear in each image in the archive, and then saving a Visual Word file for each individual image that contains a weighted list of each shape from the dictionary that appears in the image. The Visual Word files are then indexed using SOLR. Searches are conducted by retrieving or creating (in the case of a newly submitted image) a Visual Word file for the seed image and then submitting the seed file to the SOLR index where it is searched using a Bag of Words (BOW) algorithm. For those interested, the development version of the code is publicly available at https://github.com/cstahmer/archive-vision. Please note that the current codebase is very much still under development, changes daily, and should not be considered ready for release and implementation. I provide the link here so that those who are more technically minded can get a better idea of the fundamental approach that we have taken to solving the visual search problem.
LikeLike
March 25, 2013 at 7:55 pm |
Carl, many thanks for a thorough response to Anna’s and my query. I’m very much interested in how these tools work–their logic–because I believe that such knowledge can enhance how one uses a given tool and how one considers the results received.
That Arch-V is an assistive technology, a complement as opposed to a replacement for human cataloguing, seems wise. Your explanation, moreover, helps illustrate why it serves as a complement rather than a substitute for traditional cataloguing.
LikeLike
March 25, 2013 at 9:03 pm |
Thanks for this, Carl. As Eleanor observes, understanding the logic of these tools helps users better understand how to them.
LikeLike
March 31, 2013 at 4:41 pm |
[…] Other Reviews: https://earlymodernonlinebib.wordpress.com/2013/02/25/english-broadside-ballad-archive-ebba-at-ucsb/ […]
LikeLike
April 6, 2013 at 4:16 pm |
[…] English Broadside Ballad Archive (EBBA at UCSB) (earlymodernonlinebib.wordpress.com) […]
LikeLike
April 17, 2013 at 12:20 pm |
[…] English Broadside Ballad Archive (EBBA at UCSB) (earlymodernonlinebib.wordpress.com) […]
LikeLike