The Starry Night: What Van Gogh Left Out
Painted in daylight at Saint-Rémy in June 1889, it invents its village, drops the window bars, and drags a cypress close. Its sky may be real, too.
The Starry Night looks like a view copied in the dark, at speed, before it could fade. It was nothing of the kind. Vincent van Gogh painted it in daylight, in a studio downstairs from his asylum bedroom, and he edited the view hard: he left out the iron bars on his own window, invented the quiet village under the hills, and pulled a cypress tree far closer than it stood in life.
The canvas is modest in size, oil on canvas at 73.7 by 92.1 centimetres (29 by 36 1/4 inches), dated by the Museum of Modern Art in New York to mid-June 1889. MoMA, which has held it since 1941, acquired through the Lillie P. Bliss Bequest, describes it as both an exercise in observation and a clear departure from it. That tension, between what he saw and what he changed, is the whole story of the picture.
A bedroom window, a downstairs studio
Van Gogh entered the Saint-Paul-de-Mausole asylum at Saint-Rémy-de-Provence, in southern France, voluntarily on 8 May 1889, a few months after the breakdown in Arles in December 1888 in which he injured his own ear. The asylum, a former monastery, was less than half full when he arrived. That gave him room to work: a bedroom on an upper floor and a ground-floor studio. He stayed for twelve months, into 1890, and kept painting through most of it.
His bedroom window faced the countryside he would return to again and again in drawings and canvases that summer: an enclosed field, low hills beyond it, and a pre-dawn sky. In a letter to his brother Theo in early June 1889, quoted by MoMA, he set down the moment that sits behind this canvas: “This morning I saw the countryside from my window a long time before sunrise, with nothing but the morning star, which looked very big.” The morning star is the bright body just left of centre in the painting. MoMA identifies it as Venus.
On or about 18 June 1889 he wrote to Theo again, this time with a finished piece to report: at last he had a landscape with olive trees, and also a new study of a starry sky. That letter is the anchor historians have long used to date the painting to June. Note his own word for it. He called it a study, a composed work, not a sketch taken on the spot.
He could hardly have done otherwise. The vision happened at night, but, as MoMA puts it, the painting was created in several sessions during the day, under entirely different atmospheric conditions. He worked from memory, from drawings, and from watching. The sky in the picture is not a transcript. It is a reconstruction, built in working hours by a painter who had been looking hard at the real one.
What is missing, and what was moved
Start with the omission that is easiest to miss because the painting feels so open. In a letter of 23 May 1889, Van Gogh described that same outlook plainly: through the iron-barred window he could make out a square of wheat in an enclosure, and above it, in the morning, he watched the sun rise. The bars were part of the view every day. They are nowhere in The Starry Night. He painted the prospect and removed the thing that framed, and confined, his seeing of it.
The village went the other way: it was added. MoMA states directly that the village below the hills was based on other views and could not be seen from his window. He supplied it from memory rather than from Provence. MoMA notes that the church steeple resembles the steep steeples common in his native Holland, not the church towers of southern France. The calm town at the foot of the canvas, with its lit windows and straight lines, is an invention that gives the eye somewhere quiet to rest under a restless sky.
The cypress was moved, not invented. Cypresses grew around Saint-Rémy, but not where the painting puts this one. MoMA notes that the cypress at the left appears much closer than it was. In the picture it stands in the immediate foreground, dark and flame-like, rising almost to the top edge and linking ground to sky. Van Gogh enlarged it and dragged it forward, and the composition leans on that decision: without the tall dark shape at the left, the swirling sky would have nothing solid to push against.
Even the sky, the part that feels most observed, was stretched. MoMA notes that although certain features have been reconstructed as observed, Van Gogh altered celestial shapes and added a sense of glow. The sky takes up about three-quarters of the picture plane. The stars and the crescent moon at the far right are not small points of light. They are large orbs, ringed with circles of white and yellow. Van Gogh himself, writing to Theo in September 1889 about works from this period, called them exaggerations in arrangement, with lines contorted like those of ancient woodcuts. By November he was harsher still, writing to Emile Bernard about reaching for stars that were too big and calling it another failure.
One reading should be kept in its proper place. MoMA’s text suggests that, considered symbolically, the cypress could be seen as a bridge between life and death, and notes that cypresses were trees of graveyards and mourning. That is an interpretation offered by the museum, not a statement Van Gogh made about this canvas.
A new attempt to date the sky
For a long time the argument stopped there: observed in part, exaggerated in part, dated to June by the letter. In October 2026 it reopened. A team led by Elijah Jeong-hyun Kim of Kyung Hee University in South Korea published a study in Scientific Reports arguing that the Moon, Venus and the central group of stars match the pre-dawn sky over Saint-Rémy around 22 July 1889, about a month later than the usual dating.
The case rests on fitting several clues together. The central group of five stars, they argue, matches the Hyades, a V-shaped cluster in Taurus, rather than Aries, as earlier attempts proposed. The Moon is their strongest card. Around 18 June 1889 the Moon was more than half lit, not the slender crescent Van Gogh painted, and Kim told ScienceAlert that mismatch is central to the argument. Around 22 July, by their reconstruction, a waning crescent Moon sat above Venus with the Hyades in the same frame, and weather records suggest a clear enough dawn to see it.
Kim is careful about what that would prove, and the article should be too. He has said the match does not pin the completion of the painting to a single day, and that Van Gogh may have brought together the skies of several days around 22 July on one canvas. The team’s proposal is that he conceived the picture in June and may still have been completing it after 22 July.
Not everyone is persuaded. A spokesperson for the Van Gogh Museum in Amsterdam, quoted in Smithsonian Magazine’s coverage, replied that Van Gogh was not concerned with exact likenesses, so celestial positions cannot reliably date the work, and noted a serious breakdown around 22 July that left him unable to paint for about a month. Jose Luis Aragon, a physicist who has studied the painting but was not part of the new team, called the July date an intriguing hypothesis rather than a settled answer, and suggested the sky is likely a composite of many nights. His broader verdict is one both sides can share: the painting is not pure fantasy, and Van Gogh watched the real sky very closely.
Bottom line
The Starry Night is not a window left open onto a June night in Provence. It is a daytime construction from an asylum studio: a real pre-dawn Venus, a wheat field seen through bars that were then painted out, a Dutch-looking village that was never in the view, and a cypress pulled close enough to hold the canvas together. Whether its crescent Moon belongs to June, to late July, or to several nights combined is now an open, evidence-based argument. What is documented is simpler and stranger: he watched carefully, and then he changed almost everything except the watching.