Searching through archival images can be traumatic and emotional. Especially when searching in historically White institutions, photos and especially drawings depicting communities and individual people of color are often dehumanizing or even violent. What's more, erasure can often mean literal destruction of records, but it can also involve systematic lack of acknowledgement, recognition, or preservation of knowledge of a community. Photographs-- of daily life, or family portraits-- can often preserve and represent powerful moments of humanity, of daily life, of family and even of our ancestors. In my archival photographic research, a photograph of something mundane, but familiar, like a kitchen implement or a piece of clothing, would hold my attention and provide some kind of hard-to-describe connection. Whether photographs bring heavy or warm feelings, allow yourself the time to process.
For me, finding images has required luck, creativity, and persistence. Searching for "Providence Chinatown" won't bring up many photos of what the neighborhood used to look like (although this project is itself beginning to change that). Many of the glimpses we can find of the neighborhood are incidental -- in the background of photographs of other things, like the image accompanying this article about Central Baptist Church:
Images of Central Baptist Church from the Rhode Island State Archives 
"Looking in the background" is how I would describe my approach for many of the images I was able to find of Providence's Chinatown. See how in the above image, the building to the left side, which on careful inspection has an awning labeled "ORIENTAL RESTAURANT," appears again in a second image from the RI State Archives, again in the background of a photograph of Central Baptist Church, with ladders leaning against it.
The bibliography page lists a wide variety of online and offline resources, collections, and archives which I've come across, which are excellent starting points for research, especially on early Asian American communities.
I used Google and DuckDuckGo to image search by keyword, and these search terms can work for online archives as well. You may have to skim through hundreds of photos. DuckDuckGo's image search links to the full resolution image from each result, making it easier to find the original images. These search terms can also work in the Library of Congress's various search systems, such as the The Prints & Photographs Online Catalog. See Starting points on the Bibliography page for more places to search for imagery.
Starting with a phrase like "Providence Chinatown" or "historic Providence Chinatown" makes sense. You can begin to develop a list of keywords as well; mine included some of the following. Limit to a time range as well if possible.
- Chinatown
- Chinese
- Empire Street
- Walker Street (old name for Empire Street)
- Shimbo Alley
- Paine Street (old name for Shimbo Alley)
- Swanson Gangway
The last few names are of small streets and alleys that no longer exist, but were immediately around Chinatown in 1914, so they are excellent, very unique keywords which I learned from old maps. (Also see the Tracking photos section for collecting and organizing keywords and street names.) In some cities, streets can be miles long.
If there is a short street near your location, it's worth focusing on it. In New Orleans, Tulane Street is miles long, but the old name for Elk Place was South Basin Street, so "S. Basin Street" is a great search term -- and Basin Street overall is only a few blocks long.
Pair the above keywords with your place name ("Providence") or "historic" or "archival"; you can also try things like "Overview of Truckee" for photos taken from a tall building or nearby hill.
"Deadwood, [S.D.] from Mrs. Livingston's Hill" by John C. H. Grabill, between 1887-92, Library of Congress. The area of Chinatown can be seen in the distance. See Mapping photos for more on this image.
Bear in mind that not all images in collections have been digitized or are accessible online. Many are not even individually catalogued, or at some archives, catalogued at all. Visiting archives in person almost always turns up details, materials, and new clues. See Physical archives below.
Sanborn maps (see Starting points in the bibliography) were created by the Sanborn Insurance Company every few years for most cities above a certain size in the U.S. since the late 1800s. These can provide detailed context for photographs and other records, as well as addresses, details on entryways and even staircases or roofing. Searching maps before and after the period you're researching can often offer additional clues to when buildings were built or how they were modified, and when street names changed. Also see this collection of tips from the Library of Congress's Division of Geography and Maps, who have a very comprehensive collection of Sanborn maps
An 1899 Sanborn map of Providence's Chinatown, showing several Chinese American businesses. Library of Congress, (image 18).
Some local records may contain property maps and other detailed maps at a similar level of detail to Sanborn maps, and are worth searching for at local libraries, archives, and City Hall records. See Mapping photos for more on how to use this kind of map.
One way to search for images is to identify a uniquely named nearby landmark to search for, in the hopes that the neighborhood you're searching for appears in the background. Like Central Baptist Church at the start of this section, photographs of the landmark from different angles and eras can include precious glimpses of a community.
Looking over your maps and photographs, find a landmark with the neighborhood in the background, and research what names it has had over the years. In Truckee, a distinctive tower near the site of Chinatown was part of McGlashan Mansion, the home of a noted racist community leader. (See Buildings and landmarks in Tracking photos). Although ultimately I found that the tower was built after the nearby Chinatown was destroyed, learning the various nicknames the tower had over the years ("the Crystal Palace, the Rocking Stone, the Rocking Stone Tower, also known as the Ice Cone, or Monster Icicle," Truckee-Donner Historical Society) helped to widen my search of photos of the area, and later find a photo of the site before the tower was built, possibly showing a glimpse of Truckee's Chinatown district.
Building out a timeline for nearby landmarks can also help with this kind of search; see Build a rough timeline in the Tracking photos section.
The Modern Theater in 1916, showing the street corner just outside of Chinatown following its destruction in 1914, from CinemaTreasures.org 
"The great Hub-and-Hub race at Deadwood, Dak., July 4, 1888, between the only two Chinese hose teams in the United States" by John C. H. Grabill. Library of Congress
Regular events and celebrations nearby can be useful search terms; examples that helped in Providence Chinatown include Chinese New Year, while in Deadwood, South Dakota, Chinese Americans participated in town parades and competitions like the Hub-and-Hub race on July 4th, 1888, and are featured in several photographs of the event.
An interesting and powerful technique for finding additional images of a neighborhood is to use reverse image search on photographs of buildings you already have, even for the wrong time period. If you can identify a really visually unique landmark, especially a building with a unique silhouette, reverse image search can sometimes bring up other photographs of the same building. It can also sometimes surface better resolution versions of a photograph, or wider crops showing more detail around the edges, or from different time periods.
I'll post more on this technique soon.
Newspapers were, for me, the heaviest part of the research process, due to the constant use of racist language, the casual lack of respect or care for Asian Americans, and the accounts of police harassment, violence, and erasure which pervades coverage. Even present-day coverage of communities of color is disturbing, whether of historic events or not. But there was a lot of information in these archives, and I felt a responsibility to sift through them to find what I could. This sometimes meant searching every article of every issue of a newspaper for days or weeks following a major event (violence, celebrations, world events) to see what I could find. Please take a look at A note of caution, at the end of the introduction to this section, for ways to care for yourself during this kind of work.
Uniquely, many online newspaper archives include full text search, meaning that you can search within the text of the articles for keywords, just like those we used for photo searches. This also means that, unfortunately, searching for slurs and outdated or inaccurate terms can sometimes be a productive, if hurtful, search technique. Please exercise care for yourself and your community if you choose to use these. You can also refer to this list of sensitive terms and issues for Asian American histories, from the Asian American Comparative Collection at the University of Idaho.
The following list includes commonly occurring racist terms and terms used incorrectly due to racist beliefs and practices, for the purpose of archival search.
Click to view racist search terms
Chinamen/ChinamanCoolie
Oriental
Asiatic
Mongolian (for Chinese people)
Many newspapers can be found through the U.S. Newspaper Directory, a search service of the Chronicling America online collection at the Library of Congress. The U.S. Newspaper Directory has a very comprehensive list of many newspapers in the U.S. (with notable exceptions, see below), but not all of them are available for full text search in the Chronicling America search engine. Use the U.S. Newspaper Directory to identify newspapers, and Chronicling America to search within those that have been scanned. The U.S. Newspaper Directory will also list other libraries which have copies of those newspapers, and I used WorldCat.org to find additional cross-library listings. You can now also search Chronicling America by region, on a map. See the bibliography for more.
A variety of scanned pages from Providence, RI newspapers, including some with racist depictions and designs.
However, neither Chronicling America nor the U.S. Newspaper Directory has a comprehensive listing of non-English-language newspapers, nor is it currently possible to search in non-Latin characters with these tools. It is possible to search by "ethnicity press," and 57 newspapers are returned when using this to search for "Chinese" before 1920, and 19 for "Korean," although only a few of these definitely predate 1920.
I've been very interested in how early Chinese American and Asian American newspapers covered lives of people in their own communities, and have found a number of resources for in-community newspapers, which you can find in the bibliography. These come with their own challenges of access, language, and preservation, but I hope to expand this area of archival search in the coming year.
Searching physical archives is a larger topic than I can get into here, and varies widely depending on the archive. But many local and community archives are not available online. Many are open to the public; the City of Providence archives require no appointment, and contain many original photographs, as well as police records containing mugshots, property records, business directories, and marriage and court records. These can all be important parts of piecing together histories, although they focus on a narrow slice of peoples' lives. Cross-referencing peoples' names, ages, and addresses from court, marriage, deed and police records can be painstaking but revealing.
The Library of Congress has many records which are not available online, including photographs, maps, business directories, and of course, books. The AAPI Community Collection, for example, is publicly available but must be requested several days in advance and only box by box; librarians in the Asian Division can help locate specific items within, and I took careful notes on its contents myself; it contains books, manuscripts, personal records, yearbooks, maps, and a wide variety of other items. In the Prints and Photos Division, several collections are only available in person, although they are rapidly being scanned and uploaded. See the bibliography's Starting points for more specific collections to visit, and the Tracking photos section for how I keep records on what I find.
Personal, family, and community collections are, of course, the most likely to include items which can give us insight into the lives of people in otherwise erased historic communities of color. But they often exist as the result of careful private stewardship for generations, and families whose histories have been systematically excluded and erased from White institutions' collections, or whose histories have been represented in racist ways, may be very cautious about how their stories and collections are shared and to whom.
In my work in Providence's Chinatown, I have, through descendants of the community, been able to view much more personal materials, including portraits and other memorabilia. But the decision to share such items, how, and with whom, is something which requires care and is for descendants to make. In the absence of known descendants, I believe the use of family archives is something to be thoughtfully considered in dialogue with today's Chinese American and Asian American descendant communities, depending on the site and its histories.
So far, I have carefully looked through such items, and sought to incorporate understandings of their meaning into my reconstruction work indirectly. But I hope to expand on this through projects like my Ancestral Memory Enclaves workshops ), where people have an opportunity to creatively explore their connection to family photographs and memories through similar methods of relational reconstruction at a personal scale.
How many photos is "enough" to make a reconstruction? In the next section, I'll show how I make connections between photos, and how I organize them on a map to understand what I'm able to see, and what's missing. But given the amount of creativity it takes to craft a virtual reconstruction, I think it's possible to work from as little as a single photo or drawing.