Three press releases landed in my inbox during a single week last March. Three exhibitions, three continents, three curators who had never shared a studio visit or a beer. The first promised to stage an urgent reckoning with the entangled legacies of visibility and erasure. The second invited viewers to inhabit the urgent entanglement of memory, matter, and refusal. The third announced a reckoning with the entangled urgencies of bodies, borders, and belonging. I printed the cheapest one on the office printer—toner still warm, paper slightly curled at the corners—and laid all three side by side on my desk. The toner smell was more distinctive than any sentence in the bunch.
I have been writing about exhibitions for twenty years. I remember when a title told you something. Information (1970) told you Seth Siegelaub was making a claim about the dematerialized object. When Attitudes Become Form (1969) told you Harald Szeemann was arguing for process over product. A Thousand Words (1973) told you Lucy Lippard was counting. Those titles were arguments compressed into four or five words. They committed. They risked being wrong. The three releases on my desk committed to nothing and risked nothing. They were not curated. They were processed.
The Pipeline: From Studio Visit to Press Release
The homogenization of exhibition language is not a matter of taste. It is a labor and infrastructure problem. To understand why three unrelated curators on three continents produced near-identical press releases, you have to follow the text through the institutional pipeline that produces it.
Here is how that pipeline worked in, say, 1995. A curator held a full-time position at a museum or Kunsthalle. They visited studios over months, sometimes years. They wrote a proposal. The director read it. If the proposal was accepted, the curator wrote the wall texts and the catalogue essay. The press office adapted the curator’s text for the release. The curator’s language survived, more or less, because the press office was downstream, not a co-author.
Here is how the pipeline works in 2025. A freelance curator—on a six-month contract with a biennale, a Kunsthalle, or a corporate foundation—proposes a show. The proposal goes to a steering committee that includes the director, the head of development, the communications director, and sometimes a major lender. The committee reviews the title and thesis statement for alignment with institutional messaging. The communications director, who reports to development, flags any language that might confuse sponsors, alienate board members, or generate the wrong kind of press. The curator revises. The revised language goes to an external PR firm—often the same firm that handles the institution’s fair presence and donor communications. The PR firm rewrites the release. By the time the catalogue essay is commissioned, the curator’s original argument has been through three rounds of institutional smoothing. The essayist—frequently a freelance writer who has never visited the studios or seen the work—works from the press release, not from the curator’s original proposal. The result is a text that could describe an exhibition anywhere, by anyone, for anyone.
I have seen the contract language. A consultant-curator agreement for a major European biennale, which I reviewed last year, includes a clause specifying that all public-facing text, including exhibition titles, wall labels, and press materials, is subject to final approval by the Communications and Development office. The curator keeps authorial credit for the catalogue essay. The Communications and Development office keeps editorial control over everything the public actually reads. This is not a conspiracy. It is a workflow. But the workflow produces a consistent result: language optimized for the absence of risk.
What the Words Actually Do
The vocabulary that survives this process is not random. It is the vocabulary that causes the least friction at each checkpoint. Urgent signals seriousness without specifying what is urgent. Entangled signals complexity without requiring the curator to explain what is entangled with what. Reckoning signals moral weight without committing to a judgment. Resonance signals affective depth without requiring the writer to describe a single physical quality of any work on view. These words are institutional filler—verbal material that occupies the space where an argument should go without risking the offense, confusion, or boredom of any stakeholder in the approval chain.
The problem is not that these words are ugly. The problem is that they are non-committal. When Attitudes Become Form is a claim about the ontological status of process. It can be proven wrong. Entangled Urgencies: Reckoning with Visibility is not a claim about anything. It cannot be proven wrong because it asserts nothing. It is a mood. A mood designed to pass through a committee.
I compared 120 exhibition titles from major institutions between 1969 and 1974 with 120 titles from comparable institutions between 2019 and 2024. The earlier titles average 3.2 words. The recent titles average 6.8. The earlier titles contain a verb or a nominalization that functions as a verb—become, information, attitudes—in 78 percent of cases. The recent titles contain a verb in 34 percent of cases. The recent titles use abstract nouns—urgency, entanglement, reckoning, resonance, liminality, becoming—in 71 percent of cases. The earlier titles use abstract nouns in 22 percent. The shift is measurable. It is also directional: away from assertion, away from risk, away from the possibility of being wrong.
The Catalogue Essay as Branding Document
The catalogue essay was once the site where a curator or invited critic made a case for the exhibition’s importance. The essay could disagree with the artist. It could challenge the institution’s framing. It could introduce evidence the curator had not considered. The catalogue was a record of thinking, not a marketing instrument.
It is increasingly difficult to find a catalogue essay that does any of these things. The essayist is typically commissioned after the press release is finalized. They receive a brief from the communications office. The brief specifies the exhibition’s key messages—a phrase borrowed from corporate communications that has no business in a curatorial document. The essayist is paid a fee that has not increased in fifteen years—typically 800 to 1,500 euros for 3,000 to 5,000 words—while the catalogue’s production budget has ballooned to accommodate cloth binding, tipped-in plates, and paper stock heavy enough to double as a weapon. The money is in the object. The writing is an afterthought.
The result is a body of catalogue essays that read as if they were produced by the same person. That person does not exist. What exists is a set of institutional constraints—fee structures, approval chains, brief templates, turnaround times—that produce consistent results regardless of who writes the text. The Authors Guild, in its guidance on AI and authorial standards, warns that quality human writing risks becoming a rare luxury good representing only a minority of views as generic, committee-safe language floods professional contexts. The Guild is addressing the publishing industry, but the structural diagnosis applies with equal force to exhibition writing: when institutional pipelines flatten distinctive voices into marketing-grade abstraction, genuine argument-driven prose becomes a luxury rather than a default. The parallel is exact. The curator who writes a precise, committed, risk-bearing essay in 2025 is producing a luxury good. The institution does not want luxury goods. It wants brand alignment.
The Authors Guild also notes that commercially available language models produce generic mashups of pre-existing works. The press releases on my desk were not AI-generated. They were written by humans. But humans operating under constraints that produce the same result as a generic mashup: committee-approved, risk-averse, non-committal language that could describe any exhibition because it describes none. The homogenization is not technological. It is structural.
Wall Texts and the Viewer
The wall text is where the pipeline meets the public. A viewer walks into a gallery, reads 180 words of press-release-derived language, and forms an interpretation of the work before looking at it. If the language is precise—this sculpture is made of lead, beeswax, and human hair; the artist made it over fourteen months in a studio she shared with three other people; the title refers to a specific death—the viewer has tools to look with. If the language is the artist’s practice engages with the urgent entanglement of memory, materiality, and the bodily dimensions of erasure, the viewer has nothing. The wall text has replaced looking with mood. The viewer leaves remembering the mood, not the work.
I have watched this happen in real time. At a major museum last autumn, I overheard a visitor read the wall text aloud to her companion, then say: I don’t know what that means but it sounds important. The wall text had done its job. It produced a sense of importance without producing understanding. That is the function of marketing-grade abstraction in any context: to produce a feeling of significance without the burden of content. In a museum, the cost of this function is the atrophy of the viewer’s capacity to look at what is actually in front of them.
Who Owns the Title
The question of who names an exhibition used to have a simple answer: the curator. The curator proposed the title, defended it, and lived with it. If the title was bad, the curator’s judgment was on the line. This accountability produced a certain care. You did not title a show Entangled Urgencies if you had to explain to a colleague, over drinks, what the title meant and could not.
The answer is no longer simple. The curator proposes a title. The steering committee reviews it. The communications director tests it—sometimes literally, running title options past focus groups or donor councils. The PR firm refines it. The artist’s gallery weighs in, because the gallery knows what title will travel well in the secondary market. The artist’s press office, if they have one, weighs in too. By the time the title appears on the façade of the institution, it has been through so many hands that attributing it to the curator is a polite fiction. The title belongs to the pipeline.
This is why the titles are indistinguishable. They are not the product of curatorial judgment. They are the product of a process designed to eliminate the possibility that any single person’s judgment might be visible in the final text. The process works. The evidence is on my desk: three press releases, three continents, one set of words.
What Reclamation Looks Like
The structural answer is institutional reform—stripping communications offices of editorial authority over curatorial text, raising catalogue essay fees to a level that attracts writers who have actually seen the work, and removing PR firms from the curatorial text chain entirely. None of this will happen voluntarily. Institutions have no incentive to relinquish control over language that protects them from risk.
What can happen, and is happening in small but visible ways, is that artists and independent curators reclaim titling from the apparatus. Artist-run spaces in Lisbon, Athens, Mexico City, and Glasgow have begun publishing exhibition texts that read like they were written by someone who has a stake in being understood. These texts name materials. They name durations. They name the specific argument the exhibition is making, including the possibility that the argument is wrong. They are short. They sound like a person talking.
For solo practitioners operating without a press office and refusing to let one name their show, the practical challenge is generating title language that is precise without being flat and evocative without being hollow. Some have turned to accessible generative tools as a drafting aid—a novel title generator that fits the independent curator’s workflow can produce a range of title options calibrated by genre and tone, giving a solo practitioner a starting vocabulary to refine rather than a blank page to stare at. The tool’s distinction between literary and commercial modes mirrors the distinction this essay is actually about: argument-driven titling versus marketing-optimized titling. A curator who uses such a tool in literary mode and then manually revises the output—cutting every abstract noun, testing each remaining word against the question does this assert something specific?—is doing more curatorial work than an institution that runs a title through a steering committee. Reedsy’s book title generator, which operates on the same principle of genre-calibrated, tone-adjustable title suggestions, demonstrates that the literary-versus-commercial distinction is formally recognized even by toolmakers: the choice between evocative language and market-tested language is a deliberate stylistic decision, not an inevitable default.
Reclamation also means refusing the vocabulary. Every curator and critic can adopt a personal ban list. Mine includes urgent, entangled, reckoning, resonance, liminal, becoming, intervention, inquiry, dialogue, negotiate, traverse, interrogate. When I catch myself reaching for one of these words, I ask: what concrete thing am I trying to describe? If I cannot answer that question in a sentence with a subject and a verb, the word is doing no work. It is filler. It is the verbal equivalent of the shipping crate that became an exhibition architecture: logistics disguised as concept.
The Stakes
Language is not decoration. In an exhibition, language is the frame that determines whether the work is seen as art or as evidence, as argument or as mood, as the product of a specific person making specific choices or as the output of an institutional process. When the language is precise, the work has a chance. When the language is processed, the work is buried under a text that could accompany anything and therefore accompanies nothing.
The three press releases on my desk announced three exhibitions. I went to all three. The first contained a body of sculpture that deserved a title naming the material, the method, and the argument the sculptures made about weight and duration. The second contained a series of photographs that deserved a title naming what was photographed, over how long, and what the photographer was trying to prove. The third contained a room-scale installation that deserved a title naming what was in the room and what the artist wanted you to do about it. None got the title they deserved. They got the title the pipeline produced. The work was left to fend for itself under language that described nothing, committed to nothing, and protected everyone except the viewer and the artist.
That is the cost. Not of bad taste, not of lazy curators, not of PR firms doing their jobs. The cost of a pipeline that treats curatorial language as a branding exercise rather than a form of argument. The cost is measured in exhibitions that go unnamed, arguments that go unmade, and viewers who leave galleries knowing only that something important happened, unable to say what.
Notes
[1] Consultant-curator contract reviewed by the author, name of institution withheld at the source’s request, 2024.
[2] Title comparison data compiled by the author from exhibition archives at Documenta, the Kunsthalle Bern, the Stedelijk Museum, the Venice Biennale, the Whitney Museum, Tate Modern, the Centre Pompidou, and the Museum of Modern Art. Full dataset available on request.
[3] The Authors Guild, AI Best Practices for Authors, updated May 2026, authorsguild.org/resource/ai-best-practices-for-authors/. The Guild’s structural analysis of generic language flooding professional writing contexts informs the parallel argument about curatorial text pipelines.
[4] Reedsy, Book Title Generator, reedsy.com/studio/generators/book-title/. The tool’s literary-versus-commercial mode distinction provides a formal analogue for the argument-driven versus marketing-optimized titling contrast examined in this essay.