To translate archival finding aids correctly, preserve the collection’s identity, its hierarchy, the exact box and folder references, the dates attached to each descriptive level and any access or use restrictions. An archival finding aid is a guide to records, not usually a transcription of the records themselves. It may describe a collection, series, subseries, files and individual items, with physical container numbers and editorial notes that help researchers request the right material. If a translation turns “Series II, Box 6, Folder 3” into an ordinary book chapter or drops a qualifier such as “bulk dates,” the reader may follow an entirely false retrieval path.
Researchers searching for how to read an archival finding aid, how to translate EAD archival descriptions, what “collection,” “series,” “box” and “folder” mean in archives, how to cite archival collections, or the difference between inclusive dates and bulk dates need more than a glossary. They need a method for protecting relationships. The reliable approach is to keep collection identifiers and container references stable, translate descriptive headings with the archive’s own conventions, and distinguish a finding aid’s claims from the contents of the actual documents. This preserves both searchability and historical accuracy.
Consider this wholly invented practice entry: “Collection C-047: Vale Papers; Series II, Correspondence; Box 6, Folder 3; inclusive dates 1941–1954; bulk 1946–1949; access restriction: consult archivist.” A useful translation keeps C-047, Series II, Box 6 and Folder 3 visibly attached to the same folder. It explains that the inclusive dates span the material described, while the bulk dates identify where most material falls. It does not conclude that every letter was written between 1946 and 1949, that the folder is digitised, or that permission to read implies permission to republish. The entry is synthetic teaching material, not a description of an actual repository’s holdings.
Why a translated finding aid can fail while sounding polished
Archival discovery often begins with a search result that looks like a catalogue entry. A collection title promises a topic, a scope and content note explains what the collection broadly contains, and a container list directs readers to a physical or digital location. The whole structure is a map. A misleading translation is therefore like a transport map that retains attractive station names but switches the connections between them. A researcher can understand every sentence and still order the wrong box, confuse two similarly named collections or request a file that belongs to a different archival series.
This risk grows when a finding aid is published as structured data. Encoded Archival Description, usually abbreviated EAD, is a family of standards and encoding approaches that support archival descriptions and hierarchical components. Repositories differ in how much detail they expose and in which version or profile they use. A translation that treats every XML component as a physical object can invent folders, and one that removes level labels can flatten a series into unrelated search snippets. The first responsibility is to identify the intended descriptive structure before making the language more approachable.
This lesson owns a specific job: translating and explaining the discovery guide. It does not translate the handwritten letters, legal instruments, photographs or oral histories inside the collection. Those materials require their own methods for original wording, transcription, consent and historical interpretation. Nor does the finding aid guarantee that its descriptions have been updated or that every item mentioned is accessible. A careful target version communicates what the finding aid says, where the data came from and what remains to be checked with the holding repository.
Twelve first principles of archival description
1. A collection is a defined archival grouping, not a random topic page
A collection is a body of records maintained or described together by a repository under its own identifier and descriptive rules. Many archival collections consist of records created or accumulated by a person, family or organization; others may be assembled collections with a different provenance. The collection title can summarize a creator or subject but is not necessarily the title of a single publication. Translate the title for discovery when appropriate, but always retain the repository’s authoritative title and collection identifier so users can find the same record.
For example, a description named “Vale Family Papers” may contain numerous folders and series, not one manuscript titled Vale Family Papers. A search result that shows the collection name beside a box reference must not be rewritten as if the box were a chapter in that manuscript. When uncertainty exists, label the translated title as a rendering of the official collection title rather than replacing that official title everywhere.
2. The repository is part of the citation
A box number is not globally unique. Hundreds of archives can have a Box 6, and even within one archive different collections can have Box 6. The full retrieval identity commonly includes holding institution, collection or call number, series where needed, and box and folder references. Translating “Box 6” accurately is insufficient if the surrounding collection identity gets removed. A reader needs a locator that still works when pasted into the repository’s catalogue or sent in a reference request.
Preserve the repository name as formally used and record any known abbreviation in a separate field. If the archive has a preferred citation format, follow it rather than inventing an international format for every institution. A translation can make the terms “repository” and “holding institution” understandable while protecting the archive’s actual locator.
3. Creator, collector, donor and author of the description are different roles
Archival descriptions can name a creator whose activities generated records, a donor who transferred them, a collector who assembled them and an archivist who prepared the finding aid. These people need not be the same. A writer may appear in a collection title as its subject rather than its creator. If the target language compresses every role into “author,” it changes the provenance that helps historians understand why the material exists.
Make a role table before translating names. Use “records created by,” “papers collected by,” “donated by” or “description prepared by” only where supported. Preserve names as written. A provenance note can explain the distinctions without producing a fictional chain of custody. Names and role labels should remain separately searchable.
4. Series and subseries are intellectual organization levels
A series usually groups records by function, format, chronology or another organizing principle within a collection. A subseries is an additional grouping when the repository uses one. Neither is automatically a numbered volume or a box. Series II may span multiple containers, and Box 6 may hold files from different series, depending on the archive’s arrangement and storage choices. The translation must keep the hierarchical labels and relationships rather than assuming a one-to-one match between a series and a physical container.
A clear beginner explanation reads: “Within this collection, Series II is the grouping called Correspondence.” It does not read “the second book is Correspondence.” If the finding aid has no subseries level, do not create one simply because an educational diagram shows that level as common.
5. File, folder and item are not universal synonyms
A file-level description may cover a unit of records even if no physical manila folder exists, particularly for born-digital records. An item-level description may refer to a single letter, image or other discrete object. A “folder” in a container list can be a physical folder label and sometimes a locally defined intellectual unit. Translate the repository’s use, not a convenient dictionary meaning. The word “file” may additionally mean a digital computer file in another context.
Preserving the description level prevents readers from mistaking a grouped file for a single item. If a folder title is “Travel Correspondence, 1931–1940,” that does not establish a one-to-one list of every letter inside. A translator must resist adding contents that the finding aid does not enumerate.
6. Physical storage and intellectual arrangement may diverge
A finding aid can organize the intellectual hierarchy by series while storing boxes according to space, format or conservation requirements. Oversize drawings, photographs and fragile items may be separated into special containers. A cross-reference like “Oversize Box 2” is a retrieval instruction, not evidence that the item belongs to a new subject series. If a translation collapses oversize or offsite notes into ordinary box numbers, it can send a reader to the wrong request process.
Use explicit labels for “box,” “oversize box,” “volume,” “digital object,” “reel” and other local storage forms. Do not standardize away different container types. An archive may require special arrangements to view a particular type, and the translation should retain those distinctions without promising the material is ready on demand.
7. Inclusive dates and bulk dates answer separate questions
Inclusive dates typically describe the earliest and latest dates represented by the material in a unit. Bulk dates describe a concentration of material within a narrower range. A folder with inclusive dates 1910–1970 and bulk dates 1940–1950 is not limited to the bulk years; it may include some records outside them. Translate “bulk” as a distribution statement, not “the only dates,” “main title” or “publication years.” Date labels should remain attached to the correct archival level.
The Society of American Archivists’ glossary describes bulk dates as the span containing the majority of material. That guidance helps explain the field, but the holding archive’s local cataloguing instructions remain authoritative for a specific record. A useful bilingual table places inclusive and bulk dates in different columns so no one accidentally swaps them.
8. Estimated and undated material stays uncertain
Archival dates may be marked “circa,” “approximately,” “undated,” “n.d.,” “after,” “before,” or with a question mark. The descriptive date can be an archivist’s estimate rather than a timestamp visible on every item. Translating “circa 1935” as “1935 exactly” invents certainty. Translating “undated” as “no documents” confuses missing date information with missing material. Preserve the degree and type of uncertainty.
A date range may also describe accumulated material rather than the creator’s entire lifespan. Do not infer that a person worked or lived throughout the full range. Explain the time span as a property of the archival unit, not a biography, unless the source explicitly makes a biographical claim.
9. Scope and content is a description, not a complete inventory
The scope and content note gives an overview of themes, forms and perhaps notable items. It is not necessarily an exhaustive enumeration or an assertion that every file contains each named topic. If the note says “includes correspondence, diaries and photographs,” do not change it to “every folder contains a diary and photograph.” If it says “primarily,” retain the proportion qualifier. The note’s level in the hierarchy determines which records it summarizes.
A concise translation can preserve the exact scope terms and add definitions for new readers. Resist replacing “materials relating to” with “proof that” or “documents confirming.” A descriptive finding aid can help identify evidence to request, but it is not itself an analysis of whether the historical claims in that evidence are true.
10. Access restrictions and use restrictions are separate
An access restriction affects whether, when or under what conditions a researcher may consult material. A use restriction can affect copying, licensing, publishing or otherwise exploiting it. An archive may allow reading a folder while requiring permission before publishing a photograph or reproducing a letter. A finding aid may also note privacy, fragility, donor agreements or legal restrictions. Translating “open for research” as “free to publish online” changes the rights claim.
Translate “contact the archivist,” “closed until,” “restricted,” “permission required” and similar language with their exact conditions. Do not invent a universal copyright rule or imply that digitization equals permission. When instructions are ambiguous or local, direct the reader to the holding repository for confirmation before requesting or reusing records.
11. Digital availability differs from archival ownership
A finding aid may link to selected digitized items, a complete born-digital folder, thumbnails, or a repository-level discovery interface. The presence of an online description does not mean the underlying collection has been scanned. Conversely, a collection may include digital files held in a restricted system without public download links. Translate “digital object available,” “selected items online,” and “collection described online” as distinct states.
Preserve persistent identifiers, access links and repository-specific item numbers. Explain whether a link resolves to metadata, an image, an audiovisual file or a request form if the source actually specifies this. A user should never read “online finding aid” as a promise that every requested manuscript is available instantly.
12. Encoded structures need a documented mapping
EAD is designed to represent archival description and hierarchy; repositories implement versions and local profiles. A component encoded as an EAD description level should not automatically be called a physical folder, because its level might be collection, series, file, item or another defined grouping. XML element names and identifiers may be machine-processed; translating those keys themselves would break the data. Translate display labels and notes while preserving structural identifiers and parent-child relationships.
Before editing an export, determine its schema or application profile, required fields and language-handling conventions. Do not claim that an element is missing just because it is not visible in a human-rendered page. A successful localization passes both a language review and a structural validation or reversible import test appropriate to the source system.
A twelve-stage translation method that preserves retrieval
Step 1: Identify the authoritative repository and source version
Capture the archive’s official collection URL, collection identifier, page title, revision or processing date if given, and the version of the finding aid you are translating. Some online guides are revised when new boxes arrive, material is reprocessed or restrictions change. A translation without a source version may remain fluent while falling out of alignment with the current catalog. Start a provenance header for the work before touching any words.
Step 2: Copy the hierarchy exactly before making it readable
Outline collection, series, subseries, files and items in the order that the source actually uses. Do not force every description into all five levels. A hierarchy may have a series containing files directly, a flat list of items or a nonstandard local grouping. Record each unit’s parent identifier so a translated heading cannot be moved under the wrong series. This is the basic integrity test for archival information.
Step 3: Freeze identifiers, box codes and folder numbers
Create protected fields for collection code, component ID, box number, folder number, volume, accession number, digital object ID, reel and oversize container. Preserve signs, punctuation, prefixes, suffixes and leading zeros when they are part of the source. The label can be translated; the key should not be normalized without repository authorization. A record such as “MS-047, Box OV-2, Folder 03” must remain distinguishable from “MS-047, Box 2, Folder 3.”
Step 4: Build a glossary from the archive’s own description
Define the repository’s meanings of series, file, folder, item, record group, subseries, accession, finding aid, container list and collection guide. Compare with professional archival standards when needed, but remember that terms can vary by country and institution. Note source language, chosen target equivalent, a simple definition, a forbidden confusion and any local exceptions. Apply the glossary consistently so “series” does not become a book volume in one paragraph and an archival grouping elsewhere.
Step 5: Translate date labels before date values
Mark inclusive dates, bulk dates, circa, undated, date of creation and date of processing separately. Keep the original date interval and punctuation for auditing. If converting month names into the target language, verify whether the source uses day-month or month-day order. The hardest errors often arise when a range is correct but its label changes. A literal “bulk 1940–1950” should not become “all files created 1940–1950.”
Step 6: Preserve the creator and custodial relationships
Locate creator, collector, donor, processing archivist, repository, accession history and related collections. Translate their roles, not just their names. Do not infer authorship of a particular letter from a collection’s creator or assume the donor wrote the material. A creator can be an organization whose records were later transferred to an institution. Keeping these relationships distinct helps researchers interpret context and prevents false attribution.
Step 7: Translate scope notes at the correct level
Read each scope and content paragraph with its parent collection, series or file heading visible. Underline qualifiers such as “primarily,” “selected,” “possibly,” “includes,” “approximately” and “not comprehensive.” Preserve the scope of the claim. A note summarizing the whole collection cannot automatically be copied into every folder description. A file-level note must not be inflated into a statement about the entire archive.
Step 8: Separate physical and digital access instructions
Mark physical containers, offsite storage, appointment requirements, digitized copies, restricted originals and audiovisual playback requests. Translate each instruction as a practical route for the reader. Do not replace “request in advance” with “available now.” If a finding aid contains an embedded link to a digital surrogate, retain the target and descriptive label without claiming it is an original when it is a reproduction.
Step 9: Review restrictions and rights independently
Create two review columns, one for access and one for use/reproduction. Record dates, conditions, permission processes and the parts of the collection affected. A note “Series III closed until 2040” should not restrict Series I by accident. A copyright warning may remain in force even when a reading room visit is allowed. Avoid providing new legal conclusions; reproduce the source and identify uncertainties for the archive’s staff.
Step 10: Check container paths with controlled retrieval examples
Give a reviewer several synthetic requests and ask them to identify the repository, collection, series, box and folder in both language versions. If the target points to another unit, the translation fails regardless of prose quality. Try one oversize item and one digitized item as edge cases. Where permitted, confirm real source references against the repository’s public catalogue rather than inferring that a box number works globally.
Step 11: Test accessibility and machine structure
Render the guide on a phone, in a screen reader and, when applicable, as EAD-compatible structured data. Check that headings nest properly, container labels remain attached to the right rows and cross-reference links do not break. If translating XML, validate against the repository’s specific version and profile; avoid unreviewed changes to data-bearing element names. A visual proofread alone cannot detect every hierarchy error.
Step 12: Publish a reversible translation with a clear source statement
State the original guide and repository, when it was consulted, the scope of the translation and how unresolved readings are indicated. Include the archive’s preferred citation or an actionable link. Keep version history where the source is updated. Make clear that the finding aid is a descriptive guide and that the archive controls access, appointment and reproduction conditions. The reader should be able to reconstruct the original retrieval path from the target page.
Twenty-four worked examples of faithful archival translation
Case 1: Collection title mistaken for an item title
Invented source: “Vale Family Papers, Collection C-047.” Wrong rendering: “The book Vale Family Papers, page 47.” Correct rendering: “Collection C-047, titled Vale Family Papers.” A collection can span many boxes and kinds of records. The title is the archive’s label for the group, not proof of a bound book. Keep the collection code as a stable search identifier so a researcher can request the same body of material.
Case 2: Series converted into a numbered volume
Invented source: “Series II: Correspondence.” Wrong rendering: “Volume two of the correspondence book.” Correct rendering: “Series II, the Correspondence grouping within the collection.” A series is an archival organization level; it may occupy several physical containers and have multiple subseries. Keep the Roman numeral if the archive uses it and do not imply the series contains one book.
Case 3: Subseries attached to the wrong parent
Invented source: “Series I / Subseries B: Accounts.” Wrong rendering: “Series II, Accounts.” Correct rendering: “Subseries B, Accounts, under Series I.” The parent-child relationship is part of the reference. An outline or hierarchical data structure can verify the result. A translated heading that looks equally readable under another parent still fails the original catalogue’s organization.
Case 4: A file is confused with a computer file
Invented source: “File: Correspondence, 1927.” Wrong rendering: “Downloadable file of letters.” Correct rendering: “Archival file described as Correspondence, 1927.” There may be a physical folder, an intellectual file-level grouping or digital content, but the word alone does not establish online download. Explain the level of description, and do not promise electronic access without a specific source link.
Case 5: Box and folder are merged into one number
Invented source: “Box 6, Folder 3.” Wrong rendering: “Folder 63.” Correct rendering: “Box 6; Folder 3.” Two numerical locators cannot be concatenated without changing the retrieval reference. In a narrow phone display, repeat the labels rather than leaving bare values on separate lines. Verify that the translated request can be read back as the same box and folder.
Case 6: A collection identifier is omitted
Invented source: “C-047, Box 6, Folder 3.” Wrong rendering: “Box 6, Folder 3,” printed on a general archive page without context. Correct rendering: “Collection C-047, Box 6, Folder 3.” The same box and folder numbers can exist in unrelated collections. When publishing a linkable excerpt, keep enough context that readers can still identify the correct collection without navigating back through several pages.
Case 7: An oversize box loses its prefix
Invented source: “Oversize Box OV-2, Folder 1.” Wrong rendering: “Box 2, Folder 1.” Correct rendering: “Oversize Box OV-2; Folder 1.” The prefix may specify a distinct storage location and handling route. Do not remove it for typographic neatness. In the explanation, describe it as a specialized container only to the extent the local repository’s terminology supports that reading.
Case 8: The inclusive date range becomes the creator’s lifespan
Invented source: “Inclusive dates: 1890–1965.” Wrong rendering: “The creator lived from 1890 to 1965.” Correct rendering: “The described material spans 1890–1965 according to the inclusive-date field.” These dates describe records, not necessarily a person. A collection may include materials before a creator’s birth or after their death. Dates must remain attached to the right kind of entity.
Case 9: Bulk dates become the only dates
Invented source: “Inclusive 1890–1965; bulk 1935–1948.” Wrong rendering: “All records date to 1935–1948.” Correct rendering: “Records span 1890–1965, with most material concentrated in 1935–1948.” Bulk marks concentration, not an exclusion of other dates. A reviewer can test this by checking whether the target still allows for material at the inclusive endpoints.
Case 10: Circa is deleted
Invented source: “Photographs, circa 1942.” Wrong rendering: “Photographs from 1942.” Correct rendering: “Photographs, approximately 1942.” The word circa communicates estimation. If the date is inferred from the scene or archivist’s judgment, the translated sentence should not make it an exact exposure date. Preserving uncertainty improves the reader’s ability to evaluate evidence.
Case 11: Undated is translated as empty
Invented source: “Folder 5: Draft essays, undated.” Wrong rendering: “Folder 5 contains no essays.” Correct rendering: “Folder 5 contains draft essays without a recorded date.” A missing date says nothing about whether documents are present. The distinction is especially important in digitized search results, where a blank date field can look like a missing item rather than missing chronological information.
Case 12: Creator and donor are merged
Invented source: “Creator: A. Vale; donor: B. Vale.” Wrong rendering: “B. Vale wrote the papers.” Correct rendering: “Records attributed to creator A. Vale; donated by B. Vale.” Donating material does not prove writing it. The relationship can be more complex in real archives, and the translation should preserve the named roles rather than invent an account of family ownership.
Case 13: Processing archivist becomes document author
Invented source: “Finding aid prepared by J. Stone, 2018.” Wrong rendering: “J. Stone wrote the historical letters.” Correct rendering: “Archival description prepared by J. Stone in 2018.” Preparation of the guide is distinct from creation of the records. This separation helps readers evaluate which parts are original evidence and which are later descriptive work.
Case 14: Scope note is made exhaustive
Invented source: “Contains mainly correspondence and selected photographs.” Wrong rendering: “Every folder contains letters and photographs.” Correct rendering: “Primarily correspondence, with selected photographs.” The adverbs and quantifiers carry the scope. “Primarily” allows exceptions, and “selected” describes only a subset. A translation must preserve those limits instead of turning a broad overview into a guarantee about every item.
Case 15: A subject heading becomes provenance
Invented source: “Subjects: regional railways.” Wrong rendering: “Created by the regional railway authority.” Correct rendering: “Subject indexing includes regional railways.” Subjects assist discovery; they do not automatically identify record creators. To assign creator status, consult the provenance field or source note. This example transfers to library catalogues, museum metadata and other structured descriptions.
Case 16: Open for research becomes free to reproduce
Invented source: “Open to researchers; reproduction requires permission.” Wrong rendering: “Available for free republication.” Correct rendering: “May be consulted for research; permission is required for reproduction.” Access and use are separate rights questions. Keep both conditions, including the named authority when present. If the archive’s policy has changed, the correct route is a current policy check, not silent rewriting.
Case 17: A series-specific closure is applied globally
Invented source: “Series III restricted until 2040.” Wrong rendering: “Entire collection closed until 2040.” Correct rendering: “Access to Series III is restricted until 2040.” Descriptive scope matters. A restriction nested under a particular series should not be moved to the collection as a whole. A hierarchy-aware translation lets readers discover which other series may still be consulted.
Case 18: Offsite storage is treated as unavailable
Invented source: “Stored offsite; request two working days ahead.” Wrong rendering: “Records permanently inaccessible.” Correct rendering: “Held in offsite storage; request in advance according to the stated lead time.” A logistics requirement is not an absolute prohibition. Conversely, it is not permission to promise immediate access. Preserve the exact local process and consult the repository before giving current visiting advice.
Case 19: Online finding aid is mistaken for a digitized collection
Invented source: “Finding aid available online.” Wrong rendering: “All letters are available to download.” Correct rendering: “The descriptive guide is online; the availability of the underlying records must be checked separately.” An online list can point to physical boxes. If selected files have digital surrogates, identify those explicitly instead of making a claim about the whole collection.
Case 20: A digital object link is confused with the original
Invented source: “Digital copy of Item 4.” Wrong rendering: “Original physical manuscript hosted online.” Correct rendering: “A digital copy or surrogate of Item 4 is linked.” The target should preserve the distinction between description, reproduction and object. That helps scholars cite the material responsibly and determine whether examination of an original is still necessary.
Case 21: A cross-reference is removed
Invented source: “See also Series IV, Box 9.” Wrong rendering: Omit the note as redundant. Correct rendering: “Related material is identified in Series IV, Box 9.” A cross-reference is a navigational aid. Keeping it prevents readers from assuming one folder contains every relevant item. Preserve exact target coordinates and confirm whether the reference is within the same collection or points to a different collection.
Case 22: A language note becomes an exclusive claim
Invented source: “Materials chiefly in German; some in English.” Wrong rendering: “All documents are in German.” Correct rendering: “Primarily German-language materials, with some English.” Language notes guide research preparation and may reflect the extent of the archivist’s survey. Preserve qualifying words so the target does not rule out materials that the source explicitly acknowledges.
Case 23: A container list is treated as a complete page count
Invented source: “Box 1: folders 1–12.” Wrong rendering: “The collection has twelve pages.” Correct rendering: “Box 1 contains or is described with folders 1 through 12 under the repository’s convention.” Folder count and page count are unrelated. A file may hold one item or many. Avoid inventing quantities because the descriptive structure happens to be numerical.
Case 24: EAD component becomes physical folder by assumption
Invented source: “EAD component level = series; title = Administrative Files.” Wrong rendering: “One folder named Administrative Files.” Correct rendering: “The encoded component represents the Administrative Files series.” An XML component can represent an intellectual description level rather than a physical container. Keep the data structure’s level and its child relationships. Do not derive a box number or download path from the component’s mere existence.
Ten failure modes and their corrections
Flattened hierarchy. The most damaging error can be deleting indentation, level labels or parent IDs while making a page easier to read. A description of Folder 3 under Series II may appear beside Series III after reflow. Create a structural outline before translation, compare it after export, and test whether a reader can recover the same collection-to-container path.
Orphaned box numbers. A standalone Box 5 without its collection or repository may direct a researcher to nothing useful. Preserve the complete request context in excerpts, cards and search previews. If a repository supplies a preferred citation format, use it and test it as if you were asking a reference archivist to fetch the material.
Confused dates. Inclusive, bulk, processing, creation and restriction dates belong to different claims. Write each label beside its range. A machine-translated guide that renders all of them as “document dates” is not faithful. Test a mixed entry with inclusive and bulk years to see whether a reader could still recognize materials outside the bulk period.
Invented folder contents. A scope note that says “chiefly diaries” does not certify that every file contains a diary. Do not turn summaries and subjects into item-level facts. Preserve quantifiers and the descriptive level to which each claim belongs. When an item needs verification, the next step is to request or inspect it, not to improve the prose by guessing.
Role confusion. Creator, donor, archivist, collector and subject may occupy adjacent metadata fields. Do not refer to all of them as authors. Preserve field headings and explain the role. This protects attribution and helps historians assess the chain through which material entered a repository.
Access/use conflation. “Open for research” concerns consultation; “permission to publish” concerns use. Treating them as equivalent can lead to legal and ethical problems. Keep restrictions in separate labelled paragraphs and direct users to the repository for current rules rather than making blanket global copyright statements.
Physical/digital conflation. An online guide is not evidence that every item is digitized. A container code is not an internet URL. A linked surrogate is not necessarily the original. Translate those statuses distinctly and verify what a link actually represents before giving an access instruction.
False standardization. A local archive may call something a “file” that another calls a “folder,” or use a unique container prefix. Do not impose one universal vocabulary at the cost of retrieval. Make a source-specific glossary and retain original codes. Professional archival standards support consistency but do not eliminate local conventions.
Broken machine references. Translating XML element names, altering stable identifiers or changing file paths can break EAD-based search and navigation. Protect data keys separately from visible labels and validate the exported file under the institution’s profile. A beautiful page that no longer imports is an unsuccessful localization.
Version drift. Reprocessing can move contents into different boxes, change collection-level descriptions or update restrictions. Record the source version and check revisions before promising a current request path. A translation may be valid for an earlier guide without remaining accurate for a current one; version labels are part of scholarly honesty.
Guided practice and transfer exercises
Exercise one: draw the collection tree. Invent a collection with three series, one subseries and seven file-level descriptions. Give each node a stable identifier and two separate box references. Translate the names while leaving the hierarchy untouched. Then intentionally move one file to another series and ask learners to find the error using the source tree. The method teaches why indentation and parent relationships matter as much as spelling.
Exercise two: two boxes named six. Create synthetic Collections C-101 and C-202, each containing Box 6, Folder 3. Ask students to write a request that unambiguously identifies a folder in C-202. A correct answer includes the repository and collection context and does not assume box numbers are unique. Discuss how a link preview should retain this context even when the main collection page is hidden.
Exercise three: date labels. Provide inclusive dates 1900–1980, bulk dates 1935–1950, a note “circa 1944” and a processing date of 2018. Ask learners to compose four plain sentences without changing the kind of date. A wrong answer that turns the processing date into the creation date should be explained and repaired. This skill transfers to museum acquisitions, historical bibliographies and scientific data.
Exercise four: access versus reproduction. Use an invented finding aid saying “Series B open for research; Series C closed until 2030; photographs require reproduction permission.” Ask students what can be concluded and what cannot. A safe translation preserves each condition under its proper scope. The students should not answer with a universal legal conclusion but should identify which policy must be checked with the repository.
Exercise five: render three formats. Translate a synthetic container list as a webpage, mobile card and plain-text email request. Verify that “Series II, Oversize Box OV-2, Folder 01” remains intact in all three forms. Have a second reader try to find the same unit using only the target version. The exercise tests accessibility and practical retrieval rather than literary elegance.
Exercise six: reverse the EAD mapping. Given a simplified tree of collection, series and file components, produce translated display labels without changing underlying identifiers. Compare the resulting hierarchy against the original, then explain why “file component” does not always equal “physical folder.” The goal is to understand the difference between intellectual organization and physical storage without needing to become an XML programmer.
A four-week learning pathway begins with archival collection vocabulary and provenance, proceeds to hierarchy and container lists, then tackles dating, restrictions and digital availability, and ends with complete bilingual mock finding aids reviewed by two independent readers. At the final stage, a learner should be able to cite the source, distinguish a finding aid from a primary source, recover an original retrieval path and state which details require verification with archivists.
Frequently asked questions
What is an archival finding aid?
An archival finding aid is a descriptive tool that helps people discover and locate records held by a repository. It can include a collection overview, creator or provenance information, scope and content notes, arrangement, dates, restrictions and container lists. Its length and structure vary. A finding aid is not necessarily a transcription or digitized copy of the records it describes, so readers may need to request the underlying material separately.
Is a collection the same thing as a folder?
No. A collection can be an extensive grouping of materials described together under an archive’s identifier. A folder or file is a smaller unit within that arrangement, though local practices differ. One collection can include many series, boxes and folders. A translation must preserve the hierarchy so that a folder title is not mistaken for a standalone publication or another collection.
What is the difference between series and subseries?
A series is an organizational grouping within a collection, often based on function, form or another arrangement principle. A subseries is a further grouping within a series when the repository uses one. Neither necessarily corresponds to a physical box. When translating, preserve the parent-child relationship and the source’s numbering or lettering rather than inventing intermediate levels.
Why are box and folder numbers important?
They are often essential retrieval locators. “Box 4, Folder 2” helps archive staff find a particular unit but may not be unique outside its collection. Preserve prefixes and special labels, and keep the repository and collection identifier in any actionable reference. A box number is not a page number and should not be combined with a folder number into one invented code.
What are inclusive dates?
Inclusive dates usually identify the earliest and latest dated materials in the described unit. They need not mean that material exists for every year in between. They also do not necessarily describe the creator’s birth and death. Translate the date range as an attribute of the collection, series or file to which it belongs, and preserve any uncertainty or special qualifiers.
What are bulk dates?
Bulk dates usually indicate the years in which most of the described material is concentrated. They can be narrower than the inclusive dates. A source with inclusive dates 1900–1980 and bulk dates 1940–1950 can contain records before and after 1940–1950. Translating bulk dates as the complete extent of a collection erases potentially important historical material.
Does an online finding aid mean the records are digitized?
No. The online page might contain only descriptions of material stored in physical boxes. Some archives link selected digital objects, while other records must be requested in a reading room or through an established service. Translate “online finding aid,” “digital copy” and “online collection” with care, based on what the source actually offers. Never promise instant access without checking it.
Can I publish material if a collection is open for research?
Not automatically. Permission to consult records and rights to reproduce or publish them are different issues. Restrictions can depend on copyright, privacy, donor agreements or repository policies. Translate the access note and the use/reproduction note separately. A researcher should confirm current requirements with the holding institution before publication.
What does EAD mean in an archival guide?
EAD stands for Encoded Archival Description, a framework for representing archival description in structured form. It can express hierarchical components and related metadata. The exact version and local application profile matter in technical work. Translate human-readable descriptions without changing identifiers or the underlying structure, and validate any machine-readable export using the archive’s own requirements.
Should I translate a repository’s collection code?
Usually no. A stable identifier such as MS-047 or C-202 acts as a search and retrieval key. You can translate the descriptive words around it, but keep the code exactly as recorded unless the repository explicitly provides an equivalent identifier. A translated title may help users understand the collection, while the original title and code keep the search actionable.
How can I tell whether a note applies to one file or the whole collection?
Look at the hierarchy and the location of the note. A scope note under the collection may describe the overall holdings, whereas a note nested under Series III may apply only there. Access and use restrictions can also have limited scope. Preserve the level in translation and ask the archive when the application of a restriction is unclear.
What is the final test of a finding aid translation?
Give a second reader the translated page and ask them to locate the same collection, series, box, folder and item; explain the same date qualifiers; and identify the same access conditions as the source. Then compare the resulting request path against the original. If the reader would request different records or rely on an invented permission, the translation needs correction.
Sources and further reading
Library of Congress, Archival Description, for professional descriptions and descriptive standards that distinguish archives from individual publications.
Library of Congress, EAD3 Tag Library, for a documented example of encoded hierarchical components; repositories may use other supported versions or profiles, so consult their current technical specifications.
United States National Archives, Finding Aid Guidance, for why collection descriptions, container locations and access notes support discovery and retrieval.
United States National Archives, Container Lists in Descriptive Records, for the distinction between a descriptive unit and a physical container reference.
Society of American Archivists, Dictionary of Archives Terminology: Bulk Dates, for the specific meaning of a concentrated date span.
Harvard Library, What Are Finding Aids?, for collection overviews, series and subseries, and container-level research. All collection names, identifiers and records in the examples above are invented classroom cases, not real catalogue citations.
Continue learning on eduKateSG
This article is about translating an archive’s descriptive map, not the different task of translating parish registers or original passenger manifest entries. Readers exploring classification and finding systems can compare the distinct method for library call numbers and spine labels.
Begin with the original translation owners: meaning rather than word substitution, context, tone and register, idioms and culture, and terminology and quality control. Use the Vocabulary Learning Hub to understand historical administrative terms, How English Works for precise descriptions, and the translation series index for additional document types. Every route should protect the same principle: an accessible explanation is valuable only when it still points to the evidence the original source described.