The American Verse Project is an online archival database of mostly nineteenth century poetry volumes published in the USA and is part of the University of Michigan’s humanities text initiative, an extensive digital library of literary and other humanistic texts. The project provides access to the full text of 120 editions of poetry by authors who began publishing before 1921 with a focus on volumes written by women and African‐American or other poets of color (additional volumes will be forthcoming).
Volumes mentioned in standard bibliographies, literary histories, and anthologies were selected for inclusion; and faculty at the University of Michigan Department of English and the University Press assisted in the selection process. Most of the print volumes came from the University Library holdings. Using OCR scanning technology and SGML mark‐up (created in accordance with the TEI [Text Encoding Initiative] Guidelines), the project administrators have created fully searchable online texts which are available to the public.
Students and researchers interested in accessing both canonical and noncanonical American poetry can use this database to search for the occurrence of words and phrases in all the volumes included or in a selected set of texts and can browse the contents to view complete text‐only reproductions of the printed sources. In addition, the project participants are developing an online hyperbibliography of American poetry linked to the University of Michigan Library holdings as well as electronic holdings, though mostly in subscription sites.
Users have a variety of search options when they try to access either the full text of poems or to locate specific words and phrases in the volumes. Users can simply browse an alphabetical listing of the digitized volumes (by author’s last name) and access the tables of contents for each. Simple and Boolean search forms enable users to enter keywords or phrases and locate passages from different works in which they occur.
The “book‐bagged amverse” search option lets users select specific volumes for their search instead of searching the whole database. There are Boolean and proximity search features included with this option which enable users to enter more detailed searches for words, phrases, and concepts. The results are not ranked; and there is an upper limit of 100 hits; but the results give not only the passage in which the terms occur but also links to the text of the whole poem and volume in which the poem occurs. Although the various search forms do not permit users to include author(s) and title(s) in their queries except by creating a checklist in the Book‐Bagged search option, the database helps researchers perform fairly sophisticated analysis of the language of the poetry. The American Verse Project is but one of a host of free, online collections of public domain texts which have appeared in the last few years using SGML mark‐up. The University of Michigan has plans to integrate this project into a more ambitious effort to develop new technical standards for critical writing which make use of the ability of SGML to link texts to portions of other texts. Researchers in the future can retrieve spans of text without having to download them to their local machines, making critical documents more “portable” from one workstation and server to another. These experimental applications will, when fully realized in future versions of the database, make digitizing printed editions such as those selected for the project even more beneficial to literary scholars. As a free, searchable database of poems which are otherwise difficult to locate, the American Verse Project in its present form is a valuable resource for academic audiences and will be worth watching in the future as the additional experimental components are more fully developed.
