The fourth post in the blog series on the role of international collections and their collectors in promoting equity, diversity, and inclusion was contributed by Luo Zhou, Chinese Studies Librarian, International and Area Studies Department, Duke University Libraries.
Collecting for global diversity is more than a matter of identifying, locating, and acquiring distinctive, international and area studies materials for Duke University Libraries (DUL). In order for these resources to be useful to students and researchers, foreign language materials must also be described and organized in a way that makes them comprehensible, accessible, and discoverable. Librarians from the International and Area Studies (IAS) Department collaborate closely with the catalogers, archivists, and metadata specialists in our Technical Services (TS) Department in order to make that happen as smoothly and efficiently as possible with finite financial and human resources at DUL’s disposal.
As the following description of the behind-the-scenes work that went into creating multilingual metadata for the digital collection of Sidney D. Gamble Photographs demonstrates, cataloging international materials poses not only a technical and linguistic challenge, but also an intellectual and ethical one. That is because the act of translation—in this case, from one format type (analog/digital) and language (English/Chinese) to another—requires the active intervention of a diverse group of library staff with both the subject expertise and the cultural literacy to provide just the right description at just the right time.
The collection of 4,700 black-and-white photographs and 600 hand-colored lantern slides amassed by the prominent Sinologist, sociologist, and amateur photographer Sidney D. Gamble (1890-1968) depicts pre-revolutionary China’s urban and rural life, public events, architecture, religious statuary, and the countryside in the 1920s and 1930s. Over the course of his long and illustrious academic career, Gamble published seven books on China and always used his photographs to supplement his narrative. He even created an index of roll numbers, exposure numbers, and brief captions, which usually included the place names and subjects depicted in his photos. This index functioned, in effect, as the analog version of the initial metadata that was used to describe the digitized version of Gamble’s photos. But, as I will demonstrate below, this was only the beginning of the process of cataloging, contextualizing, and providing access to the images in the Gamble digital collection.
As Duke’s subject specialist on China, I started working on the black-and-white images in early 2008, when the Gamble collection first arrived at DUL’s Rare Book and Special Collections (now Rubenstein) Library. The original nitrate negatives had just been digitized and placed in cold storage, to preserve them in perpetuity. Gamble’s own handwritten and typed captions, which were digitized alongside these fragile negatives, were transformed into raw text using optical character recognition (OCR) software. The digital version of Gamble’s captions thus became the foundation for the image captions and geographic headings of the Gamble digital collection as a whole. The collection was published on the DUL website in fall 2008 and immediately attracted the attention of researchers worldwide. The hand-colored lantern slides were digitized and added to the digital collection in 2014. Another photo album, containing 170 images of Gamble’s first China trip with his family, was the latest item digitized and added to the database in 2019. The Gamble digital collection that now resides on DUL’s website and servers, therefore, is a careful compilation and comprehensive presentation of all his photographs and slides together with metadata in four different languages: English, Chinese, Japanese, and Korean.
Creating bilingual geographic headings in English and Chinese
In 2008, the expert staff in DUL’s Technical Services department extracted out of the raw text from the digitized image labels a list of all toponyms (proper names of places) identified by Gamble. Removing the duplicates left a list of roughly 1,000 alphabetically organized entries for me to work on. That is, I took on the task of adding the Romanized form of Chinese characters—a system of transliteration known as Pinyin—to these place names. After going through the spreadsheet and identifying (and excluding) duplicates, I was able to reduce the total list to about 500 relatively unique geographic place names.
Next, I grouped these Chinese place names into three general categories. The first category consisted of the proper names of well-known and popular places that foreigners regularly visited in early twentieth century China. Gamble’s spellings of these place names—such as Hangchow (for Hangzhou 杭州) and Beijing’s Lama Temple (for Yonghegong 雍和宫)—relied on an earlier, popular, Romanization form of Chinese characters (known as Wade-Giles) and was relatively easy to identify for anyone familiar with the history of the transliteration systems used in the field of Chinese studies. The second, and much smaller group of Chinese toponyms, consisted of photographs depicting locations in neighboring Russia, Japan, and Korea; places with general, descriptive titles (“On the Sea”); and those that lacked any identifying information. The geographical heading assigned to this group usually just referred to the names of the countries that Gamble had visited on his various trips to Asia. The third, and most challenging, group consisted of toponyms for remote or lesser-known locations, as well as those known by a different geographical name than the one in use today. Gamble’s Romanization of these place names was inconsistent and often did not use the standard systems available to him. In fact, many of the geographical names in the last group could not properly identified until a year or two after Gamble’s photos arrived at Duke, sometime in 2009 or 2010, when I was able to do additional archival research about Gamble’s trips and his work in China. And a few were identified more than six years after the database was published when a Chinese blogger provided a clue.
Initially the digitized images and the typed image labels resided in separate locations on the DUL server. Since these items were not yet linked to each other in the database, it was impossible to compare Gamble’s photographs with the captions and geographic locations that supposedly described them. Luckily, after inspecting the physical materials, I realized that Gamble filed his negatives with roll number and exposure number in the order of his visits to different places. By arranging the photos by their roll numbers, I was able to reconstruct his trips in sequence. Inspecting the physical collection also led me to conclude that Gamble used two different photo cameras, which he called “Camera A” and “Camera B.” The negatives of the photos produced by these cameras have roll numbers from 2A to 95A and from 1B to 77B. Later Gamble relied primarily on Camera A, so we have roll numbers from 96 to 663, which actually are 96A to 663A.
To give you an example of how I used these archival discoveries to improve the metadata for the Chinese place names used by Gamble, let’s examine the images in rolls from 2A to 95A and from 1B to 36B, which cover the places Gamble visited from May to October in 1917, when he arrived at Shanghai, before travelling up the Yangtze River into Sichuan (四川) Province. Since most of the photos from this trip were taken in Sichuan Province, the place name that Gamble assigned to Image 1 (Fu Chou in roll 21A) must be located somewhere in Sichuan, despite the fact that this place name also sounds very similar to FooChow (Fuzhou 福州), a city in Fujian (福建), which is an entirely different province. So it is reasonable to conclude that Gamble’s designation (Fu Chou) actually refers to Fuzhou (涪州), which later came to be called Fuling (涪陵), a city famous for its pickles (榨菜). In order to make this location more discoverable in the digital collection, the metadata for this geographic place name now includes both its modern name (Fuling) and its old name (Fuzhou).
To take another, somewhat more complicated example, let’s inspect the toponyms that Gamble assigned to the photo of the walking “spinner” (Image 2). This label confusingly refers to two different place names: Li Fan and Tsa Ka Lo. However, since this photo comes from roll B22, these places must also be located in Sichuan Province. Li Fan is clearly a reference to Lifan County (Lifan Xian 理番县), which changed its name to Li County (Li Xian 理县) in 1945. The County is in today’s Aba Tibetan and Qiang Autonomous Prefecture (阿坝藏族羌族自治州). The Chinese character 番 means “foreign tribe” while 理 means “to manage.” People living in this region are mainly Tibetans and Qiangs, plus some Hui and Han Chinese. In the eighteenth century, Qianlong Emperor appointed rotating officials to rule this region as a way of incorporating minority groups living on the frontiers of the Qing Empire. The Chinese character for “foreign tribe” (番) was removed from the county name during the Republican era (1912-1949) because of its derogatory connotation, suggesting that the residents of this region were not Han Chinese. The county government sits at Zagunao Town (杂谷脑镇), which comes from Tibetan phrase for “land of good fortune,” as heard and spelled by the Han Chinese. It is not difficult to match Zagunao with Tsa Ka Lo, Gamble’s Romanization of the Chinese place name. The metadata for this place now includes both the old and the modern names (in Pinyin and Chinese characters), which makes it easier for users of this digital collection to match the image with other, textual sources.
Locating the actual place name for Gamble’s “So Village” presented a somewhat different challenge. That name appears on 85 photographs in rolls from both Camera A (44A to 51A) and Camera B (17B to 19B). Judging by its location in the sequence of photographs, this village must also have been located in Sichuan Province, most likely somewhere in the triangle region formed by Mao County (茂县), Wenchuan County (汶川县) and Li County (理县). Since there are also at least three different Chinese words for “village”—cun (村), zhuang (庄) or zhai (寨)—locating it in one of these counties required figuring out what Gamble meant by the word “So.” My initial guess, which was based on the assumption that “So” referred to the name of the honored ancestor of one of the more prominent families in this village, led me to suggest that Gamble’s “So Village” was really called either Suo Cun (索村) or Su Cun (苏村). Unfortunately, neither place name was found in that geographical area.
The actual name and location of Gamble’s “So Village” remained a mystery until a researcher in Sichuan brought my attention to a blog post from China in 2014. According to the Chinese blogger, the name of this agricultural settlement was the Village of Suo Chieftain (Suo Tusi 索土司), called Wasi Tusi Guanzhai (瓦寺土司官寨). With this hint, I went through DUL’s copy of the local gazetteer for the counties of Mao and Wenchuan and confirmed the blogger’s findings. The village of So is located in Wenchuan County and is populated mainly by Rgyalrong Tibetans, who moved to this part of China a very long time ago. Rgyalrong Tibetans believe that they are the descendants of the mythological “Great Peng Bird” (Dapengniao, 大鹏鸟) and therefore use a bird as their totem. Interestingly, one of Gamble’s photos (Image 3) shows a guardian statue on top of a gateway: it has a bird’s head and a human’s body and is holding a snake in its hands. This image from the Gamble collection matches the description of the village entrance in a local gazetteer published in 1997. And so, now, the metadata for “So Village” has been updated to read: Wasi Tusi Guanzhai (瓦寺土司官寨), located in Miansi Town (绵虒镇) of Wenchaung County (汶川县).
Adding Metadata, Adding Value
As these three examples suggest, identifying and assigning accurate geographical descriptions to the photos in DUL’s Gamble digital collection is as much an art as a science. Usually, it depends on a knowledge of the language and history of China and a good bit of research. But, sometimes, all you need is a helpful hint from a user located on the other side of the globe. The fact that the blog post about the contemporary name of “So Village” would never have been brought to my attention if the digital collection did not include Chinese language metadata only serves to confirm the importance of creating bilingual geographic headings for digital collections of non-English materials.
The “added value” of revising metadata cannot be put into strictly monetary terms; nor can it always be counted, like the number of clicks on a webpage. The value of including bilingual geographic headings, for example, is not merely a matter of convenience, i.e. the fact that it allows researchers who may be familiar with only one place name to identify all the photographs in the Gamble collection that are associated with this toponym. Revision of metadata also makes it possible to uncover the existence of suppressed, unknown, or undocumented subjects in DUL’s image collections (such as the Han Chinese of Zagunao and the Rgyalrong Tibetans of Wasi Tusi Guanzhai). But its true value lies in the intention to establish a meaningful connection between the international and area studies materials collected by DUL’s subject specialists and the researchers who use them, irrespective of where these researchers happen to live, what religion they practice, and what language they speak. And it is precisely because of DUL’s commitment to such cosmopolitan values that the geographical headings in the Gamble digital collection will continue to be updated and revised as new research findings come out.