The Emergence of Advertising in America: 1850-1920is a digitized collection of 3300 items related to the early history of advertising in America from the David M. Rubinstein Rare Book and Manuscript Library at Duke University. Browsing or searching the collection is straightforward and easy to do, and the metadata information facilitates searching by categories that are specific to this particular collection of objects, such as searching by company or by product. But the metadata is also somewhat inconsistent between objects and there are missed opportunities to describe items more fully or to offer contextual information for them. Moreover, there is no metadata information about the digitized objects themselves: all the available metadata refers only to the original objects.
Advertisements in the collection are described by up to eleven metadata categories, although not all categories appear for every object. The eleven categories are:
- Rights Note
- Name of Digital Collection
These categories can be somewhat confusing. For example, title and headline are nearly always the same for ads. For example, an 1860 advertisementfor a slave auction in Charleston, South Carolina is described by the first lines of the advertisement: “A Prime Gang Of 158 Negroes. By Louis D. DeSaussure.” It’s not clear why title and headline are two different categories given that they are typically identical. This also seems a missed opportunity for a more descriptive title, such as Advertisement for the auction of 158 slaves by Louis D. DeSaussure, Charleston (1860).
The description category is underutilized in this database. Some objects do not include a description category at all for no apparent reason. For others, the metadata in the description category is minimalistic and doesn’t include commonly used terms that could make searching the database easier. The entire description for the ad described above, for example, is “list of slaves’ ages and abilities.” The description field could include some of all of the professions of the enslaved people listed, as well as the names of the enslaved who are being auctioned. But more importantly for many searches, the description could include terms that would be useful for people searching for items on slavery. The word “plantation” appears in the ad, but nowhere in the metadata. While this is clearly a slave sale or auction, the word auction is not used in the metadata. A search for “auction” returns only those sales of slaves that are described in their own title as “by public auction” because that information then appears in the title field. The very minimal description here (and for other items) could be augmented by contextual information, in this case related to the extent of the domestic slave trade, to the centrality of Charleston as a site for slave trading, or biographical information about Louis DeSaussure.
The “Company” metadata category seems to be used here as a replacement for the more common “Creator” category, and in some cases that makes sense. The company is straightforward for some products (Kodak, Ford, etc). But for other kinds of ads, it isn’t clear how the identity of the company was determined. In the ad for the sale for 158 Negroes, the company is listed as T. Bennett Lucas, presumably because the ad identifies the enslaved people as being sold on his behalf. But in other cases of slave sales where the identity of the original owner is not mentioned, the slave trader is listed as the Company in the metadata.
The “Subject” description of the metadata is useful, but can also be quirky and limiting. Perhaps because the description category is so thin in this database, the Subject category seems to be used in an expansive way. The subject African/African-American is used in the case of ads that include a single picture of an African-American or when any of the content in any ways relates to African Americans. A 16-page booklet 1882 booklet advertising a “Mastodon Show” is described with the subject headings of African/African-American and Expositions/Fairs/and Festivals. Why African/African-American has been singled out as a subject heading is unclear, especially when there are no subject headings for Indians (a group named in the booklet), dogs, or circus animals. This item has no description metadata category at all, but a meaningful description could make this particular item discoverable in any number of ways that it is not discoverable now. In other cases, common terms could be added to Subject headings to better facilitate searching. The ad for the 1860 sale of 158 enslaved people referenced above is listed under the subjects of “Children, African/African-American, Slaves, but not under the term “slavery.” A search of the database for the term “slave” returns 13 items, but searching for “slavery” finds nothing.
Perhaps the biggest missed opportunity is the failure of the metadata for ads to describe their location, although metadata for other kinds of materials (still images of billboards, for example) do include a location field. Many of these advertisements are from specific locations which are specified in the text, so it’s unclear why they would not include a location field. The failure to describe the ads in this way limits the capacity to search by place. A search for Charleston thus only returns items where Charleston is part of the title or the company name (The Charleston Hotel; Charleston Teapot Company). The several ads in the collection for slave auctions in Charleston do not appear in a search for Charleston or for South Carolina.
The metadata for The Emergence of Advertising in America collection facilitate some kinds of searches while limiting others. One could easily trace the evolution of advertisements by the Kodak company by exploring the 557 Kodak ads in date order. The database makes it easy to search by product type (cigarettes, for example) or by the publication where an ad appeared (so a user could explore the different products advertised in the Ladies Home Journal). But the lack of location information and descriptions for many items make it harder to use this database to do geographic searches or to find information that might not appear in a title. Searching for “dogs,” for example, brings up 11 items which include “dog” in the description field. But it does not return the booklet for the Mastodon Show, which included among its acts a dog circus featuring canine comedians. What other dog comedians might lurk among these items? We might have to look at all 3000 objects to know.