VMS LEADERSHIP — Describing Visual Materials: A Community of Practice in Progress
By members of the Description Section and Visual Materials Section.
The VMS Steering Committee extends a heartfelt thank you to everyone who joined us for the VMS Meet-Up during Archives*Records 2026. It was wonderful to gather with colleagues and members of our community in New Orleans, to reconnect in person, and to share conversation and camaraderie at Salon Salon after a busy conference day.
We are also grateful to the in-person and virtual facilitators, presenters, hosts, and participants who helped make the VMS and Description Section co-sponsored session, Describing Visual Materials: A Community of Practice in Progress, such a thoughtful and engaging conversation. To serve as a continuing resource, we have put together the following summary, which offers an overview of the session’s structure, major themes, and key points raised by participants, along with links to the Padlet discussion boards as we continue building a shared community of practice around visual materials description.
At the 2026 SAA Annual Conference, the Description Section and Visual Materials Section co-sponsored Session 704, “Describing Visual Materials: A Community of Practice in Progress.” Paige Adair of The Portman Archives chaired the session, with a hosting panel that included Aliza Leventhal of the Rhode Island School of Design; Jacob Wolf of the Getty Research Institute; Sabrina Gorse of Heritage Works, Inc.; and Maggie McCready of the Library of Congress Prints & Photographs Division (of the Visual Materials Section Steering Committee), as well as Betts Coup and Max Goldberg of Harvard University; and Sarah Bush of Washington University in St. Louis (of the Description Section).
Designed as an informal, hybrid conversation rather than a series of formal presentations, the session brought participants in the room and online together to consider the tensions among value, representation, and resources in visual materials description. A pre-conference poll helped shape the session, while two Padlet boards served as shared spaces for note-taking and virtual participation, as a workaround to system constraints for online participants to engage. The in-person participants broke up into five groups, with a host moderating the discussion of each question, coming back together with their online peers after fifteen minutes of conversation.
The session began from the premise that description is a series of choices shaped by time, context, and capacity. The scale and research value of a body of visual materials may point toward one level of description, while the representational context of the materials, the needs of particular communities, or the risks associated with access may point toward another. Budget, staffing, technical infrastructure, and institutional priorities further influence what repositories can accomplish and sustain. Rather than attempting to establish a universal model, the session invited archivists to compare how they navigate these choices and to share approaches that make visual materials accessible without losing sight of inclusivity, care, or the realities of available resources.
To see the Padlet boards directly, please click on each question below.
Responses to the first question immediately complicated the meaning of “accessible.” Participants identified several overlapping but distinct forms of access: whether materials are discoverable by their intended audiences; whether users can intellectually understand what they have found; whether the physical or digital formats can actually be consulted; and whether the descriptive systems and digital objects are accessible to people who use assistive technologies. A collection may be technically open while remaining difficult to find, and an image may be available online while offering little meaningful access to someone who cannot see it. Participants therefore resisted treating digitization, online publication, or the existence of a finding aid as sufficient evidence that visual materials are accessible.
Much of the discussion focused on the contextual and subject information needed to make images interpretable. Photographs and other visual records often do not “announce” their purpose or function in the same way that a letter, report, or other textual record might. Traditional elements such as title, date, and creator remain important, but users may also need information about what is depicted, the circumstances in which an image was created, its physical or digital format, and its relationship to other materials. Understanding the terminology and practices of the creating community may be especially important. When materials cannot be digitized, physical description and contextual information may shoulder even more of the burden of allowing users to determine whether an item will meet their needs.
At the same time, some of the information most useful for discovering visual materials—particularly date, creator, and subject—may be difficult or impossible to establish with certainty. Participants discussed the uncomfortable role of informed guesswork: an archivist may be able to determine that one image predates another or recognize a probable subject without being able to assign a precise date or identification. Metadata lost before transfer can be especially difficult to reconstruct from an image alone. The conversation suggested a need for transparent ways to express approximation and uncertainty rather than either omitting useful information or presenting inference as fact. Internal repository notes can also preserve observations, decisions, and institutional knowledge that may not be appropriate for public-facing description but could help future staff revisit or improve the record.
The most sustained thread concerned whether visual materials require item-level description. Item-level records can give users the format, version, title, date, and content information they need to choose among photographs, recordings, or other materials. They are also extremely time-consuming to produce; moreover, archival description by its very nature is intended to be in the aggregate. Participants noted that extensive item-level description can sometimes make discovery more confusing rather than less: a search may return numerous records that appear distinct but represent duplicate prints, multiple negatives of the same image, or slightly different views of a single subject.
Aggregate description was not seen as inherently inadequate simply because the materials were visual. Instead, participants discussed a tiered or extensible approach in which a sustainable baseline creates initial access and more granular work is added where research value, user demand, risk, format, or representational context warrants it. Thematic groupings, searchable keywords, controlled vocabularies, representative images, and relational links among creators, projects, formats, and intellectual content were all raised as ways to bridge the distance between broad aggregate records and comprehensive item-level description. Guidelines based on recurring categories of material or use cases could also help archivists make consistent decisions without applying the same level of work to every visual object.
Participants also emphasized that the descriptive requirements of visual materials cannot be separated from their physical and technical characteristics. One intellectual object may consist of several physical components, while a digital visual record may depend on particular software, hardware, or an emulated environment. Different versions or views may need to be intellectually linked even when they are stored or represented separately. Whether materials are physical or digital, and whether users can directly view them or require staff and equipment to mediate access, may shape what information is absolutely necessary in the description.
Appraisal was a term that came up both in the room and online in conversation. Before a repository accepts a large body of photographic negatives, slides, complex design records, or digital visual materials, it should consider whether it can realistically make those materials accessible. The total cost of stewardship encompasses much more than storage: it may include arrangement, description, preservation, rights research, digitization or rendering, technical maintenance, and the ongoing labor of delivering the materials to users. Participants described human capacity as a major obstacle in large image collections, particularly when appraisal before accessioning has not sufficiently accounted for the labor required to identify or differentiate thousands of similar items. Selective digitization and representative samples were discussed as ways to create a useful visual preview alongside initial aggregate description without suggesting that every object can or should be digitized.
The question also moved beyond archival discovery and delivery to disability access. Participants pointed to alt text, long descriptions, audio description, accurate transcription of inscriptions, and systems compatible with screen readers and other assistive technologies. They shared web accessibility guidance and emphasized that an effective description does not have to become an exhaustive visual essay: it should convey the information needed to understand the purpose and content of the image. Questions regarding the use of artificial intelligence (AI) came up in this portion of the conversation: AI might assist with batch production of preliminary alt text for large photographic collections, but the resulting descriptions would still require human verification. AI might change the distribution of the labor, in other words, but would not eliminate the need for time, judgment, or accountability.
Participants suggested that accessibility begins with defining the intended audiences and forms of use, providing sufficient subject and contextual information for discovery, ensuring that the material and its descriptive interface can actually be used, and making appraisal and descriptive decisions that the repository has the capacity to sustain.
The second question highlighted the immediacy of visual materials. A reader can pause before continuing through a textual description, but an image may confront the viewer before they have had an opportunity to understand its context or decide whether they wish to engage with it. Once seen, it cannot be “unseen.” This distinction heightens the importance of where descriptive context appears and whether a digital interface inadvertently exposes users to violent, demeaning, private, or otherwise distressing imagery through search results and thumbnails.
Participants discussed a range of approaches to content warnings, without reaching a consensus that one practice would be appropriate in every case. Some saw value in direct, prominent notice about potentially offensive language or imagery. Others were hesitant to place archivists in the position of determining what a user will experience as harmful and preferred specific, contextual description (DACS-compliant scope and contents or historical notes) that allows users to make their own informed choices. General warnings may also flatten meaningful distinctions among different kinds of material, suggesting a need for more nuanced language and more flexible descriptive fields than many current systems provide. Some participants argued that clear, DACS-aligned description could perform much of the work often assigned to a generalized content warning, while others saw warnings and description as complementary. A few participants mentioned the use of a controlled vocabulary for types of sensitive materials. (E.g., nudity, death, etc.) For these participants, these controlled vocabularies were unique to their specific archive. Archivists created a list of terminology based on common types of sensitive materials in the archive and then attached content warnings to the record at the collection-, box-, or folder-level when cataloging.
There was greater convergence around designing access so that engagement with an image is not forced. Suggestions included blurred thumbnails accompanied by an explanation, click-through or interstitial screens, a contextual cover page preceding a digitized image, and descriptive information presented before the image itself. These possibilities also exposed technical constraints: repositories may want a pop-up, opt-in mechanism, or differentiated fields for sensitive content but find that their discovery system or digital asset management platform cannot support it. The principles established in descriptive guidelines therefore have to be translated into concrete interface and system requirements if they are to shape the user’s actual encounter with the materials.
Digitization itself emerged as an ethical and descriptive decision. Selecting a potentially offensive image for digitization or prominent display may give it greater visibility than other parts of a collection. Conversely, declining to digitize it may suppress evidence that is historically significant or important to the communities represented. Participants asked when such images should be prioritized, what context should accompany them, and how users can be offered a genuine opportunity to opt into viewing them rather than encountering them unexpectedly.
Privacy and security presented related but distinct concerns. Visual records may reveal information whose sensitivity is less apparent in accompanying textual records. An architectural floor plan, map of a protected archaeological site, photograph of a patient, or image of critical infrastructure may require different access controls than a letter or report documenting the same subject. In some cases, repositories need staff and researchers to know that relevant records exist without publicly advertising the details of materials that only a limited group may consult. This makes description inseparable from decisions about mediated access, restricted metadata, user authentication, and the institutional or legal frameworks that determine who may see what.
Consent was an especially difficult part of this discussion. Photographs may document people who never expected the images to enter an archive or circulate globally online. Historical collections frequently lack releases or other documentation establishing what kind of consent was given, how an image could be used, or whether that permission had an expiration date. Photographs of children, patients, and other vulnerable subjects make these uncertainties particularly consequential. Even images made at public events raise questions about whether attendance constituted meaningful consent to future archival use. Participants noted that expectations surrounding photography and online circulation have changed substantially, making it difficult to project contemporary assumptions backward or to determine what the people depicted understood at the time. In many cases, the answer may simply be unknowable.
Inclusive and reparative description also becomes especially complicated when harmful language is embedded in an original caption, title, or other creator-supplied metadata. Retaining that language without explanation may reproduce harm, while silently replacing it may obscure the history of the record, sever evidence of the creator’s perspective, or reduce discoverability for researchers searching the terminology used in the original context. Participants emphasized the importance of clearly distinguishing transcribed or creator-supplied language from archivist-supplied description. Contextual and processing notes can document what was retained, changed, or supplemented and explain the rationale for those interventions.
The same need for context applies to images containing symbols, terminology, or representations whose meanings have changed over time. As historical events become more distant, users may be less likely to recognize the circumstances in which an image was created, while something included in an image (symbols, etc.) may acquire new associations that shape its contemporary reception. Participants also raised the importance of applying identity description consistently rather than treating whiteness, maleness, or other dominant identities as unmarked defaults, while avoiding unsupported assumptions about race, gender, or other characteristics based on appearance alone. Controlled vocabularies can support consistent discovery, but they cannot substitute for collection-specific judgement or explanation.
Also, participants discussed ethical conflicts surrounding archival materials that may embarrass or reflect poorly on an institution. Sensitivity in this context is not just user-based but also concerns the institution an archivist is working for. In corporate and academic settings, higher-ups may not be enthusiastic about highlighting archival materials that portray an institution in a bad light. However, recognizing this history is important to not silence the realities of the past. Participants discussed the challenges associated with this, since challenging one’s employer could lead to retaliation, which may result in archivists losing their jobs.
Finally, the conversation emphasized that sensitivity is contextual and cannot be determined solely from the object or by the archivist working in isolation. The people and communities closest to the materials may understand their potential effects, histories, and appropriate presentation differently from repository staff or general audiences. Participants recommended consulting subject specialists and community expertise where possible and conducting additional research where such consultation is unavailable. They also stressed the value of documenting descriptive choices so that future archivists can understand and reconsider them as terminology, technologies, users, and social contexts continue to change.
To wrap up the session, audience members were given the opportunity to ask their questions of the group, as well as share thoughts and insights. It was a dynamic conversation for the final ten minutes of the session, making for a positive conclusion both in the room and online. The Description Section and Visual Materials Section plan to revisit this topic in further conversation, as there is certainly excitement and engagement with it, and a true interest in developing a community of practice.
Comments