© Copyright 2026 Mastersinlibraryscience.org.
How librarians built a major research collection through grey literature and hard ethics calls
Archival ethics now sits at the center of collection building rather than at its margin. In October 2026, The Harvard Crimson reported that Harvard's Fairbank Center collection began with books stashed in a bathtub in Coolidge Hall, then an apartment building, before Nancy Hearst became its first full-time librarian.
Her acquisition work, from anonymous boxes to roadside salvaged posters, is the case study: provenance gaps, unsolicited donations, and cross-border sourcing. The practical question for LIS students is less whether such collections are legitimate than how cataloging, China's 30-year rule, and access decisions determine what survives.
Major research collections are rarely the product of a collection-development policy. More often they are the residue of individual relationships, and the Fairbank Center's Chinese holdings at Harvard's Fung Library are the clearest recent case.
Nancy Hearst arrived as the first full-time librarian for the Fairbank Center for Chinese Studies Collection, inheriting an archive that consisted largely of books stashed in a bathtub in Coolidge Hall, then an apartment building. What she built over the following decades became a research collection of real depth on the Cultural Revolution and modern China. The founding instruction came from a historian rather than a strategic plan: Roderick MacFarquhar, then director of the Fairbank Center, told her to "go and just see if you can buy books."
MacFarquhar was not a passive administrator. He wrote the trilogy "The Origins of the Cultural Revolution" and taught a General Education course, Foreign Studies 48: "The Cultural Revolution," that enrolled 900 students. His research and teaching created demand for material no vendor was reliably supplying. After his death, part of his personal collection was donated to the Fung Library, closing the loop: a scholar's working library became an institutional holding.
For LIS students, the lesson is uncomfortable and useful. Large collections are frequently assembled through improvisation and personal networks, not policy, and the bill comes due later. Whoever inherits that material has to reconstruct provenance, impose order on boxes that arrived unnamed, and decide what the collection actually documents. Hearst's achievement is real; so is the descriptive work it left behind.
Approval plans and standing orders assume a stable publishing market, a known vendor, and a title you can preview before it ships. The acquisitions work that built Harvard's Cultural Revolution holdings assumed none of those things.
Grey literature means material produced outside commercial publishing: pamphlets, internal reports, posters, mimeographed circulars. Ephemeral materials are items nobody expected to survive. A revolution generates both in abundance, and vendor catalogs miss both almost entirely.
Nancy Hearst bought books in China by weighing footnotes and tables of contents under time pressure. There were no reviews to consult and no second chance to order the title next fiscal year. Selection became disciplined guesswork.
None of these routes fits neatly into a collection development policy. Anonymous donors, salvage from trash, and cross-border document acquisition raise provenance, legal, and ethical questions that standard policies rarely anticipate. The tension is genuine: the same routes that supply irreplaceable material also ask librarians not to inquire too closely into how it arrived.
Provenance is the documented chain of ownership and custody that lets a researcher say where an item came from and why it can be trusted. The provenance problem begins when that chain is missing. Nancy Hearst described the routine at Harvard's Fung Library: "One day I'd come in, and there'd be a big box from an unknown." An unmarked box is a routine accession decision with no donor to interview.
Professional standards do not demand that every undocumented item be refused. The Society of American Archivists' Code of Ethics calls for professional judgment in acquisition and for documenting collection decisions, so the archivist's role in shaping the record stays visible. SAA reappraisal guidance notes that unknown provenance is common and that its absence does not by itself block reappraisal or deaccessioning1. Options include suspending the item until its status is clear, checking title questions with counsel, documenting the process, and retaining donor communications1. ALA and SAA access guidance treats donor anonymity as valid only when restrictions are reasonable, documented, and time-limited, not as an unconditional right2.
Unknown provenance weakens evidentiary value: an item without a custody record is harder to cite as evidence of what circulated, where and when during the 1966 to 1976 Cultural Revolution. Accepting materials without asking how they left their country of origin creates institutional risk, from title disputes to diplomatic exposure3. The practical response is an accession record that logs "donor unknown" with the circumstances, interview notes, and a chain-of-custody note. When no donor exists to sign a deed of gift, a repository still has to decide under time pressure: keep the item with a documented rationale, return it, or suspend it, and record that rationale so a future archivist can defend the call.
The material you need is usually held somewhere other than where you would first look, and often in another country. Use this table to route a request before you fill out an interlibrary loan form: Red Guard serials and ephemera point one way, internal Party documents another, and Chinese-language monographs a third. Where a source did not establish digitization or access terms, the cell says so rather than guessing.
| Repository | Location | What It Holds | Digitized Online? | How to Access |
|---|---|---|---|---|
| Harvard-Yenching Library, Harvard University | Cambridge, Massachusetts | Chinese-language general, archival, special, and rare collections for modern Chinese history; holds the New Collection of Red Guard Publications | Partial: many special collections digitized through CURIOSity and HOLLIS Images; archival coverage is not comprehensive | HOLLIS finding aids; reading-room appointment and advance request; remote research accommodated only for material already digitized or handled through paid on-demand digitization and local research services |
| Fung Library, Fairbank Center for Chinese Studies, Harvard University | Cambridge, Massachusetts | Fairbank Center collection built from 1980s buying trips, unsolicited donations, big-character posters salvaged from roadside debris and provincial markets, and Roderick MacFarquhar's personal library | Not established in the sources reviewed | Onsite at Harvard; catalog records through HOLLIS; remote use depends on item-level digitization |
| Hoover Institution Library and Archives, Stanford University | Stanford, California | Chinese Cultural Revolution Collection of Red Guard newspapers, pamphlets, broadsides, and flyers | Partial: catalog records and finding aids online, and a digital portal covering posters, photographs, manuscripts, moving images, and sound recordings; full-text coverage of this collection is not established | SearchWorks and Online Archive of California finding aids; physical items consulted through Hoover and Stanford research facilities |
| Wilson Center History and Public Policy Program | Washington, D.C. | Cold War-era international history documents, including translated Chinese government materials in its digital archive | Digital archive format; much of the material is available online | Free public access through the program's website |
| Library of Congress | Washington, D.C. | Chinese-language holdings served through the Asian Reading Room; Cultural Revolution-era items are not itemized in the sources reviewed | Selected items in the Library's digital collections | Onsite reading room; interlibrary loan through participating libraries; some material online |
| Chinese Cultural Revolution Database, Chinese University of Hong Kong | Hong Kong | Chinese-language primary documents from the 1966 to 1976 period, compiled as a research database | Digital by design; remote availability not established in the sources reviewed | Typically reached through an institutional or subscription license; confirm terms with your library before planning a visit |
Censorship in a digital archive rarely announces itself. A journal issue that turned up in a search last semester now returns nothing; a catalog record loses a subject heading; a document set quietly shrinks between two visits. For anyone researching the Cultural Revolution (1966 to 1976), the working question is how to tell a genuine restriction from an ordinary access failure, and how to record what you saw so someone else can check it.
Exclusive licensing can place a journal in one database only, so absence from CNKI does not by itself prove removal. Log the date, URL, catalog change, and any screenshot. Treat unexplained absences as open questions rather than findings.
Two things decide whether a Cultural Revolution file reaches a reading room: the statutory closure period, and the discretion of the institution holding it. Only the first is written down.
For decades China's Archives Law opened state-held records after 30 years. The 2020 revision, adopted in June 2020 and in force January 1, 2021, shortened the general period to 25 years from creation for repositories at or above county level, and permitted earlier release of records in the economy, education, science and technology, and culture. Education was a new addition.
Continued closure past 25 years is allowed for archives involving state security, vital state interests, or those "unsuitable for disclosure." The law sets no fixed maximum. Access measures fall to the National Archives Administration, subject to State Council approval; a single local repository does not decide on its own.
Cultural Revolution material from 1966 to 1976 is past both the 30-year and the 25-year marks. The statute is not what keeps it closed. Classification, exemption claims and institutional judgment are, and the date on a document says little about the stamp on its file.
Repositories must publish catalogues of open records, and the law gives users a complaint right when access is refused; failure to provide access is treated as a violation. None of this guarantees what a researcher actually receives at the desk. The catalogue and the complaint route give users leverage; they do not settle a classification dispute.
The Fung Library and Harvard-Yenching collections were not assembled by people following a collection-development manual. They were built by librarians who could read the material, spot its value inside an unmarked box, and judge the risk of accepting it. Four competencies recur in this niche.
Working with Cultural Revolution material means moving between traditional and simplified characters and across romanization systems, so a single title may appear in the catalog under Wade-Giles, under Hanyu Pinyin, and under a transliteration nobody standardized. An archivist who cannot recognize all three variants will report an item as missing when it is already on the shelf.
Subject expertise is a hiring criterion here, not a bonus. You need to know what a Red Guard serial is, why a provincial imprint carries different evidentiary weight than a Beijing one, and which document series were internally classified at the time of publication. Ma Xiao-He, the Chinese Collection librarian at Harvard-Yenching, works with the Service Center for Chinese Publications because he can evaluate what the vendor offers.
That vendor relationship runs on trust and anonymity. Knowing what not to ask, and what not to write down, is part of the job.
Export controls, institutional counsel, and donor confidentiality all shape what a library can accept. Write down your reasoning, because the next archivist may have to defend the acquisition long after you have left.
Archival work rarely makes anyone rich, but the pay is steady and the ceiling is higher than the median suggests.
Archival ethics has moved from an appendix of professional codes to a standalone unit in most MLIS programs, but the teaching still centers on principles rather than pressure. The Harvard case shows where that split matters.
Representative offerings sketch the standard map. Michigan's SI 580 treats ownership, access rights, justified restrictions, donor relations, privacy, and legal obligations as ethics topics1, and SI 632 pairs appraisal and acquisition with policy questions through a decolonizing lens.2 UCLA's Archival Studies specialization builds policy development and memory-making into its core3, while Information Studies 262B ties provenance to stewardship and intellectual property.4 Simmons lists LIS 438 Introduction to Archival Theory and Practice in its 2026-2027 catalog within an Archives Management concentration.5
Most of what happened at the Fung Library fits neatly into existing units: collection development policy, deeds of gift and donor confidentiality, and the access-versus-risk tradeoff when an item arrives without paperwork. An unmarked box from an unknown donor is a deed-of-gift problem with the signature missing.
The shortfall is rehearsal. Students practice appraisal repeatedly, sorting folders against retention schedules, but rarely practice the decision Hearst described, a box appearing one day and a judgment owed by afternoon. Ethics gets discussed; it does not get timed.
A workable fix is a simulated accession memo: one anonymously donated item, an ambiguous chain of custody, a deadline, and a written recommendation covering provenance, restriction, and disclosure. Grade the reasoning, not the verdict.
A Cultural Revolution research request is won or lost in the catalog, not in the reading room.
Start in HOLLIS with broad English keywords, then rerun the query using the exact subject terms from the strongest record you find, such as the Cultural Revolution heading for China, 1966 to 1976, or China politics and government, 1949 to 19761. Add a place, an institution, a person, or a source type to narrow the set. Then search the Chinese characters directly: 文化大革命, 红卫兵, 大字报, and the traditional forms such as 紅衛兵. The romanization guess changes your hit rate, because North American catalogs follow Library of Congress conventions rather than one tidy pinyin standard. Test wen hua ge ming against wenhua geming, Mao Zedong against Mao Tse-tung, Beijing against Peking.3 In WorldCat, run characters first, then pinyin, then English.2 A character search that returns nothing proves nothing, since many records carry only romanized titles or English headings.
A catalog record says a collection exists. A finding aid says whether the series you need is processed. Read the scope and content note, then the series arrangement, and note whether the material is described at box level or listed only as an unprocessed accession.
Holdings do not guarantee scanning, lending, or admission. Identify the repository and the exact series, box, call number, or reel, then email the subject librarian with that citation in both English and Chinese4. Offer to begin with a small pilot batch. Where digitization is not offered, ask about interlibrary loan, digital surrogates, or an authorized local researcher.4 Log your access date in your own citations.
