III½. The Delft school and the rivals

A town's optics

Vermeer did not invent the luminous interior alone; he perfected it inside a local school with a specific, datable project. Around 1650 Delft painting swerved, almost collectively, towards problems of light and space. Gerard Houckgeest and Emanuel de Witte began painting the town's two great churches from oblique angles, sunlight striping the whitewashed piers and the tomb of William the Silent — perspective made devotional, the national shrine as a study in daylight. Carel Fabritius arrived from Rembrandt's studio and reversed his master's poles: instead of figures spotlit against darkness, dark figures against radiant plaster — the Goldfinch's wall is the first Vermeer wall — and experimented with wide-angle perspective boxes and murals. Paulus Potter passed through; Jan Steen kept a brewery in the town in the 1650s; Egbert van der Poel made a career of painting the 1654 explosion over and over. This was the environment — half workshop tradition, half optical laboratory — in which the young Vermeer turned from Italianate history to the lit room, and the turn looks less like a private miracle when the neighbours are visible: what Vermeer added to the Delft project was not light but stillness — the removal, one by one, of everything in the scene that moved.

De Hooch, ter Borch, and the division of the spoils

The closest comparison, then and now, is Pieter de Hooch, who worked in Delft from about 1655 to 1661 — precisely the years of Vermeer's conversion to the interior — and whose courtyards and rooms share the tiled floors, the open doors, the light-through-doorways, so exactly that priority has been argued in both directions (the sane answer: exchange, with de Hooch probably first into the sunlit courtyard and Vermeer transforming what he borrowed). Set side by side, the two painters divide the world between them: de Hooch is a painter of space — his rooms open into further rooms, his light travels, his people are furniture of the perspective; Vermeer is a painter of presence — his space closes around a single figure, his light arrives rather than travels, and the perspective serves the person. De Hooch left for Amsterdam and grander, colder rooms; his late work declines as Vermeer's concentrates, and the pairing became art history's standard lesson in the difference between an excellent painter and something else. Gerard ter Borch, older and Hague-based, supplied the repertoire itself — the satin skirt, the letter, the toilette, the parting soldier: nearly every Vermeer subject exists first as a ter Borch — and the single document linking them (both signed the same notarial act in Delft in 1653, weeks after Vermeer's wedding) suggests real acquaintance. Again the transformation is the point: ter Borch's satin scenes are exquisite anecdotes — glances, negotiations, innuendo; Vermeer takes the same room and stops the clock. Metsu in Amsterdam answered Vermeer directly (his Dublin pair of letter-pictures quotes the motifs and even, some think, the manner); Frans van Mieris and the Leiden fijnschilders pushed finish where Vermeer pushed tone. In the great 1696 auction the Vermeers and the de Hoochs were priced in the same range; it took two hundred years for the difference to become obvious, and the difference — presence against anecdote — is what every later account of him is trying to name.

The subjects: letters, music, pearls, maids

Seen as a set, the thirty-six pictures sort into a handful of obsessively repeated themes, and the repetitions are structural, not lazy. Letters (six pictures): the letter is the interior's window to the invisible — love, absence, news, the war, the sea — and Vermeer's letter-women are the purest cases of his withheld narrative: we see the effect, never the cause; the Dresden Girl Reading opens the series, the Woman in Blue is its summit, the Dublin Lady Writing with her Maid its late, layered close. Music (eight pictures): music was the era's polite code for courtship — duets, lessons, the offered lute — and Vermeer's music pictures run the code's whole register, from the Procuress's paid company through the Music Lesson's charged decorum to the late London pair, where the standing woman's Cupid promises open love and the seated woman's Baburen Procuress on the wall shadows it with purchase: the two pictures, almost certainly conceived as pendants, are his last word on the theme — love's music, bought and free. Pearls (a dozen pictures): the pearl is Vermeer's favourite object because it is his art in miniature — a sphere that is nothing but reflection, a highlight made thing; the Woman with a Pearl Necklace worships light with one; the Girl's earring is two brushstrokes, the most famous highlight in art. Maids and mistresses: the Milkmaid alone gives the working woman the monument, but the late pictures pair mistress and maid twice — the maid as messenger, witness, patience personified — a quiet sociology of the household he lived in, crowded with women (wife, mother-in-law, daughters, servants) whose names the archives keep. And behind them all, one absence: no violence, no drunkenness after 1657, almost no men except as visitors, no streets after 1660, no death. The censored world is the composed one; what Vermeer left out of Dutch painting is as deliberate as what he kept.

A geography of the œuvre

Because the œuvre is so small, its present geography is itself a map of collecting history, and the works page that accompanies this essay can be read as an atlas. The Netherlands keeps seven: the Rijksmuseum's four (Milkmaid, Little Street, Woman in Blue, Love Letter) and the Mauritshuis's three (View of Delft, Girl with a Pearl Earring, Diana). America holds the largest share — the Metropolitan's five, the Frick's three, Washington's four (one now "attributed"), the Gardner's stolen Concert, the Leiden Collection's small Virginal — the harvest of the plutocratic decades, 1890–1930, when Europe sold and the Gilded Age bought. London has three (the National Gallery pair and Kenwood's Guitar Player), the Royal Collection one, Dublin, Edinburgh, Paris (two), Frankfurt, Berlin (two), Dresden (two), Brunswick and Vienna one or two each — and every one of these placements has a story the catalogue tells: royal Saxon purchase, Rothschild bequest, wartime seizure, national subscription. Nothing has entered the canon by discovery since the 1920s except by attribution war; nothing will. The finitude is now part of the experience: seeing all the Vermeers is a completable pilgrimage — people undertake it, book by book and city by city — and the 2023 Amsterdam gathering of twenty-eight was the pilgrimage performed collectively, once, for everyone at the same time.

The household and the models

One question the crowds always ask has a better answer than the novels give: who are the women? The pictures use, over twenty years, a small repertory of faces and one wardrobe — the yellow satin jacket trimmed with ermine-like fur appears in six pictures and in the death inventory; the pearl earrings and necklace recur; the same green skirt, the same red bodice. The house on the Oude Langendijk was full of candidates: Catharina, bearing children through precisely the years her supposed likenesses were painted (the Woman in Blue's silhouette has been read as pregnancy since the nineteenth century — Dutch fashion's box-pleated jackets make it undecidable, which has not stopped anyone); the eldest daughters, Maria and Elisabeth, adolescents by the mid-1660s, whose ages fit the tronie girls and the late virginal players so neatly that whole attribution theories have been built on them (one recurrent speculation makes a daughter not model but painter of the weakest late pictures; scholarship declines it, novelists decline to); the maids — the household employed at least one, and the fictional promotion of the Girl with a Pearl Earring to a maid named Griet is Chevalier's invention, not history's. The honest position is strict: not one Vermeer sitter is documented, not even for the two pictures that look most like commissioned portraits. What the archives do give is the density of the household itself — at its fullest, the painter, his wife, his mother-in-law, up to eleven children and a maid, in one canal house with one painting room — and the inventory's poignant tour of it: the children's beds, the worn chairs, the paintings by other masters stacked as stock, and upstairs "the yellow satin mantle," two easels, three palettes. The rooms of the pictures are that house with the noise painted out; the women of the pictures are that family made anonymous — intimacy generalised into image, which is why every generation since has been able to move in.

Two women deserve to step out of the anonymity, because the documents give them character. Maria Thins, the mother-in-law, emerges from thirty years of notarial paper as the household's iron: managing rents and lawsuits, sheltering her daughter from a violent brother, lending the couple money and forgiving it, owning the Baburen Procuress that her son-in-law painted into two pictures — a Catholic gentlewoman who took a tavern-keeper's convert son into her house and, on the evidence, backed his unmarketable slowness for two decades; without her there is no Vermeer œuvre, and she has no portrait. And Catharina: her petitions after 1675 — negotiating with creditors, pleading the war's ruin, trying to keep the Art of Painting out of the bankruptcy by deeding it to her mother — are the only surviving prose from inside the painter's world, and their mixture of grief, shrewdness and exhaustion is the closest thing we have to his voice's echo. She lived until 1688, poor, moving between relatives with eleven children; whatever the pictures owe her face, they owe her endurance more.

The laboratory Vermeer: three case studies

No painter's surfaces have been interrogated harder in the last decade, and three campaigns show what the instruments now add to the eye.

The Girl in the lab (Mauritshuis, 2018). The Girl with a Pearl Earring was examined for two weeks in a glass enclosure on the museum floor — macro-X-ray fluorescence, reflectance imaging, digital microscopy, paint-sample analysis — with the public watching, and the findings redrew the picture without moving a molecule. The blank dark background is not blank: it was painted as a glossy green curtain, diagonal folds and all, whose indigo and weld glazes have degraded to near-black — the Girl originally turned before a hanging, in a shallow space, not in an abstract void; the void is chemistry, and part of her modern mystique is a pigment's mortality. She has eyelashes, invisible until the scans found their fine strokes; a signature top left; and the pearl is confirmed as almost nothing — two strokes of lead white over a shadowed ground, no hook, no outline: the most famous jewel in art is an event of paint. The ultramarine of the turban came from Afghan lapis of the highest grade; the light entered, as always, from the left, and the underlayers show the composition set down wet and sure, without revision.

The Cupid uncovered (Dresden, 2017–21). Conservators removing degraded varnish from the Girl Reading a Letter at an Open Window confirmed what X-rays had suggested since 1979 — a standing Cupid on the rear wall — and established, by studying the dirt layer between paint and overpaint, that the covering was applied decades after Vermeer's death: not the artist's revision, as scholarship had comfortably assumed, but another hand's taste. The uncovering split opinion in public — the emptied wall had become, over two centuries, part of what "Vermeer" meant, and its filling felt to some like the loss of a masterpiece of restraint — and the episode is now the standard classroom case for the questions restoration cannot dodge: whose picture is it, the author's first intention or the object history made? Dresden chose the author; the debate chose neither, and continues.

The Milkmaid's changes (Rijksmuseum, 2022–23). Scanned for the 2023 exhibition, the Milkmaid yielded an underlayer with a jug-rack and a fire-basket behind the maid — painted out by Vermeer himself in favour of the bare wall — and swift, confident first sketching beneath the figure. Here the subtraction was the artist's: the same instrument that returned Dresden's Cupid certified Amsterdam's emptiness, and the pair of results, published within two years, define the method's honesty — the laboratory does not flatter the legend; it tells each canvas's own story, and the stories differ. What the campaigns share is a revised Vermeer: not the effortless transcriber of appearances, but a fast, decisive underpainter who then slowed to a crawl in the finishing — invention quick, perfection patient — and a colourist whose economics (lapis everywhere, even under shadow) presupposed a patron who paid for what no one would consciously see.

Perspective, thread and pinhole

One technical discovery deserves its own section because it settles an old caricature. In the 1980s conservators found, at the vanishing point of more than a dozen Vermeer interiors, a tiny hole in the paint layer — the mark of a pin, from which, the paint ridges show, a chalked string was stretched to rule the orthogonals of tiles, window-bars and virginal-lids directly onto the canvas. The finding matters twice over. It proves the perspectives were constructed, not traced from any projection — a camera obscura image cannot be pinned and strung — and so retires the crude version of the optics thesis in which Vermeer is a copyist of screens; and it locates the constructions' cunning: the pinholes sit exactly where meaning wants them — at the Balance-holder's hand, between the Music Lesson's couple, at the writing woman's eye — so that the room's geometry converges, invisibly, on the picture's heart. Instrument and geometry, lens and thread, describe a double practice: he studied how lenses render the world (the halations, the scale-jumps, the melting near-field are lens facts, chosen and kept), and he built his spaces with the draughtsman's oldest tools. The caricature — genius as gadget — collapses into something better: a painter who used every way of knowing appearance his century offered, optical, geometric and manual, and subordinated all three to composition. Delft's twin exports of the decade, one might say, were Leeuwenhoek's animalcules and Vermeer's daylight: the same lens culture, aimed inward and outward.

Colour as structure

A last technical point, because it underwrites everything the criticism says about his "harmony." Vermeer organises colour architecturally, not decoratively. Each mature picture is governed by one interval — nearly always blue against yellow, the cool of daylight against the warm of things — stated in the largest areas (jacket against wall, skirt against tablecloth) and then echoed downward in diminishing amounts to the smallest accents; red is admitted as punctuation (a ribbon, lips, the Red Hat spent whole); black and white are reserved for the geometry (tiles, frames, window-leads) that keys the rest. The underlayers work for this system invisibly — cool grounds beneath warm passages and warm beneath cool, so that every surface vibrates slightly against its own base; green earth cools the flesh shadows (a Vermeer fingerprint the forgers missed); and natural ultramarine, mixed into greys, whites and even browns, tunes the whole room to sky. The result is measurable — colorimetric studies find his value-range narrower and his hue-relations more consistent than any Dutch contemporary's — and visible to anyone: it is why a Vermeer announces itself across a gallery before the subject can be read, as a chord announces a key. Cézanne's dictum that when colour is at its richest, form is at its fullest, has no better proof a century early; and the moderns who claimed him — from the Impressionists to Hammershøi's grey rooms to Hopper's lit corners — were claiming, precisely, the discovery that a picture's emotion can be built in its intervals, before and beneath its story.