How to Write an Art Review From Encounter to Supported Judgement

A review that matters moves from direct encounter to a judgement the writer can defend. You are not recording a spontaneous impression. You are distinguishing what you saw from what you think about what you saw, and making that distinction visible to a reader who was not in the room. The sequence is not a formula. It is a discipline that prevents the review from collapsing into memoir on one side or press release on the other.

Start with the encounter. Walk the gallery without a notebook. Let the hang do its work. The curatorial position is in the sequence of rooms, the height of the works, the colour of the walls, and the company a painting keeps. Beginning by reading the wall text has reversed the order of evidence. The wall text is a position. The hang is a position. The object is primary. Spend at least 20 minutes in the space before you write a word.

The review has four parts. Encounter, description, analysis, judgement. Encounter is the unedited physical experience: the light, the sound of the air conditioning, the distance you stood from the work. Description is what you saw. Analysis is what the work means in a context. Judgement is whether it succeeds, and by what standard.

This structure works for a museum show review, a single-work notice, or a long-form reconsideration. The proportions shift. A short newspaper review spends half its length on description. A journal piece spends a third on analysis. But the sequence holds. A reader cannot trust a judgement that arrives before the evidence.

Encounter

Describe the conditions. A painting seen at Frick Madison in 2023, stripped of the Frick’s domestic interiors and hung in the Breuer building’s travertine galleries, read differently from the same work seen at 1 East 70th Street before the renovation. The temporary severing of paintings from domestic architecture made the curatorial position unusually legible. Note the display rotation interval. A work that has been on the wall for 18 months is a work the institution is arguing for. A work that appears for six weeks and vanishes into storage is a footnote.

Description

This is ekphrasis, the verbal record of a visual object. Record what is there. Colour, scale, surface, material, condition. If the work is at the Yale Center for British Art, note the linen wall covering and the oak floor. The architecture is part of the account. Do not interpret yet. A sentence such as “the canvas is unlined, with a horizontal crack at the lower left quadrant” is description. A sentence such as “the painting radiates melancholy” is judgement, smuggled in early. Keep them apart.

Analysis

Analysis brings in context. The exhibition record is the permanent scholarly document that outlives the show, and you must consult it. It tells you what the curator thinks the work means, what the provenance is, and what preservation philosophy has been applied. A minimal-intervention policy leaves a tear visible. A restorative approach makes the object look as the maker intended. The choice is a position, and the catalogue should state it.

Contextual analysis places the work in a career, a period, a material tradition. Formal analysis attends to line, colour, composition, and the internal logic of the object. Both belong here. Neither is the last word.

Judgement

Judgement names a criterion and applies it. “This show succeeds because it shows a painter discarding a signature style and failing, which is more instructive than a greatest-hits show.” The reader can disagree with the criterion, but they can see what it is. A supported judgement is not a score. It is a position a reader can argue with, because the writer has shown their work.

The confusion that sinks most reviews is the conflation of description and judgement. Description records what the reviewer saw. Judgement evaluates it against a stated criterion. The sentence “the surface is thickly impastoed, with brush hairs embedded in the paint” is description. The sentence “the impasto is self-indulgent” is judgement. A reader can accept the first and reject the second if the review keeps them separate. If the review blends them into “the paint is slathered on,” the reader trusts neither.

Name your criterion early. A review of a Met Breuer commission series piece might judge the work on whether it answered the building. A review of a Soane Museum room might judge it on domestic interpretability, the quality that makes a work legible when lived with rather than photographed for a catalogue. A white-cube gallery photograph misleads because it strips the context the artist assumed. The same work in a domestic interior, with glazing that reflects the fire, reads as a different object. Your criterion makes the difference legible.

Avoid the period room misreading. A preserved interior such as the Soane Museum is an accretion of multiple periods, not a single historical snapshot. Soane’s own “Crude Hints” manuscript imagines the house as a future ruin, which means the instability is built in. The 1833 Act of Parliament froze the arrangement, but light damage and textile fading mean a significant proportion of surface area has been replaced per half-century. When you describe a room, you are describing a preservation history, not a moment. Say so.

Before You Write

A gallery review method that works under deadline has three stages. First, walk the show without stopping. Time it. A 12-room show takes 40 minutes at a steady pace. Note where you slowed down and where you sped up. The hang is working on you before you work on it. Second, walk it again with a notebook. Write description only. No adjectives of quality. Third, sit in the last room or the café and write the first sentence of the review. If that sentence is judgement, delete it and write a description sentence instead.

Verifying the Facts

Check the exhibition record before you file. The publication cycle varies. A show that closed six months ago may have no catalogue yet. A catalogue published three years after the close is a different genre of object, closer to a Festschrift than a record. Note the lag. If the catalogue is available, check your facts: the date, the medium, the dimensions. A wall text can misstate the medium. A catalogue essay can misstate the provenance. The Paul Mellon Centre for Studies in British Art sets the evidential standard for this kind of object-based art history. Its monographs are the model. Your review should approach that level of care, even at a tight word count.

Reading the Rotation

The Sir John Ritblat Gallery rotation at the British Library is a case in point. The selection policy is a position about what a national collection owes the public continuously versus occasionally. A reviewer who does not know the rotation interval cannot judge whether a particular object is being proposed as a permanent treasure or a temporary loan. Ask the curator. The answer is part of the review.

Architecture Is Never Neutral

A museum show review must account for the architecture. The white-cube gallery is not neutral. It is a 20th-century argument about attention, and it strips all contextual information except the work and the wall. The Breuer building at Frick Madison was the opposite: a heavy, material building that refused to disappear. The travertine, the concrete, the trapezoidal windows all competed with the paintings. The curatorial position became about that competition. A show that worked at the Breuer might fail at a neutral space, and the review must say which.

Domestic Spaces, Living Arguments

The Morgan Library & Museum is a different problem. The room is the argument about what a cultivated interior owes a city. A reviewer cannot treat the Morgan as a white cube with nicer wallpaper. The domestic scale, the bookcases, the ceiling height all shape what can be shown and how it reads. The same is true of the London Library, where open stacks and minimal weeding mean the collection is a sedimented record of member interests since 1841. A show in that space is an intervention in a living argument, not a display in a void.

Conservation as Evidence

Check the preservation philosophy before you write. A minimal-intervention policy at the Yale Center for British Art means the building's material ageing (travertine, linen, oak) is part of the interpretive problem. A reviewer who complains about a stain without knowing the policy has missed the point. The stain is a document. Describe it as one.

Element	What It Does	Common Failure
Ekphrasis	Verbal record of visual object; records scale, surface, condition, material	Smuggles judgement into description: 'sloppy brushwork' instead of 'visible brush hairs'
Formal Analysis	Attends to line, colour, composition, internal logic of the object	Recites schoolroom terms without connecting them to the specific work
Contextual Analysis	Places work in career, period, material tradition, exhibition history	Relies on wall text or press release rather than exhibition catalogue
Supported Judgement	Names a criterion and applies it; makes the grounds of evaluation visible	Pronounces without stating the standard: 'this is a masterpiece'
Exhibition Catalogue	Permanent scholarly record that outlives the show; source for provenance and preservation data	Cites a catalogue that is out of date or published years after the close
Gallery Hang	Curatorial position made through sequence, height, wall colour, adjacency	Reviewer treats the hang as neutral background rather than part of the work

The Period Room Misreading

The period room misreading is the most persistent error in writing about historic interiors. The Soane Museum is not a snapshot of 1837. It is an accretion of Soane's lifetime, his descendants, two world wars, and a century of preservation interventions. The "Crude Hints" manuscript is the foundational document for reading preservation as a creative act, not a freezing. When you write about a period room, state the date of the arrangement you are describing and the date of the most recent treatment. If you cannot find those dates, your account is incomplete.

The Reprint Is Not the Object

The "definitive edition" trap catches writers who cite a reissue as authoritative without checking the editorial apparatus. A Persephone Books reprint in a uniform grey jacket is a position about a neglected text. It is not a facsimile. The paper, the binding, the margins are all 21st-century decisions. A Criterion Collection edition with a spine number is a canon-making object. The supplementary materials shape the interpretation. If you are writing about the film, cite the edition you watched and note its limitations. A 35mm print screened at Film Forum is a different object from a 4K restoration streamed at home. The projection format is part of the evidence.

Spine-Fade and Circulation

Spine-fade misattribution is a material fact that corrupts judgement. A row of faded spines on a domestic shelf can reflect sun exposure on a single wall, not a reader's priorities. The London Library's open stacks show this at scale. A book that looks unread may have circulated 40 times and been rebound. Binding durability, measured in estimated circulations before rebinding, is a better index of use than spine condition. If you are writing about a personal collection, look inside the books.

The single error that sinks more reviews than any technical mistake is reviewing the idea of the show rather than the show itself. A writer reads the press release, skims the catalogue, looks at a few installation shots, and writes a review that could have been filed without entering the building. The review is coherent, well-structured, and worthless. It describes a curatorial intention that the hang failed to deliver, or it praises a work that was hung so high the surface was illegible. The encounter is the evidence. Skip it, and the review is fiction.

A second error is failing to distinguish the object from its reproduction. A facsimile silently corrects printing defects, adjusts contrast, and crops margins. A digital projection of a scanned 35mm print is not a photochemical screening, even if the cinema advertises it as one. The Criterion Collection edition of a film is not the film as projected at BFI Southbank in 1973. The difference matters. State what you saw, in what format, under what conditions. If you do not, the reader has no reason to trust the rest.