How Paris Paintings Are Dated

When letters are not enough, the works themselves provide converging evidence

The date of a painting is rarely read directly from the picture. This is especially true of the Paris period: Van Gogh often worked continuously, the works were unsigned, and a letter did not always give a clear title on the day it was written. A reliable date is therefore not an oracle, but an interval reached by several strands of evidence converging.

The first line: letters, but not just one sentence

Letters can provide a place, the weather, what Van Gogh was doing, and the order in which certain works were sent or completed. But a single letter usually illuminates only one part of the scene. To place it on a timeline, it must also be compared with the work's title, the pigments mentioned, seasonal subjects and the next letter.

Letter numbers must also be checked against the standard edition. This essay uses the project's existing let numbers only as navigation; it does not turn sentences from an older translation directly into front-end quotations.

The second line: seasons and the city

Flowers, leaves and light are clues to the season. Paris streets, the Seine and the view from Theo's apartment are spatial clues. Whether a scene could be seen from a particular window, and whether the direction of the riverbank fits the city's terrain, both affect the dating.

View from Theo's Apartment corresponds to cat. 95, F341/JH1242. It is not JH1100; that number belongs to cat. 56d, F265. Separating the two numbers is the first step in returning a “Paris landscape” to a specific place.

The third line: canvas and pigments

The support, ground and combination of pigments change with a period of work. They cannot provide an exact date on their own, but they can help establish whether two works are close in time, or whether one painting is earlier than another group. The dimensions and condition of the canvas edges may also provide clues to sequence.

This kind of evidence does not need to be written as an excessively technical laboratory report. The conclusion that matters to a reader is simpler: a date comes from several layers of a work, not only from what its subject appears to be.

Covered images and continuity of touch

Hidden images beneath an old canvas show that the support passed through different stages. Repeated contours, short strokes and colour juxtapositions between neighbouring works resemble the fingerprint of a working habit. Together, they can place a painting within a more credible interval of production.

Seeing the method through three numbers

• The quinces and lemons in cat. 126b are F602/JH1343; the fruit composition in cat. 128 is F383/JH1339; the dish with citrus fruit in cat. 89 is F340/JH1239.
• By the Seine in cat. 107 is F299/JH1254, not F269 as the OCR error suggested.
• The women in cat. 53 and cat. 54 remain under attribution review; where there is no JH number, one should not be invented for the sake of completeness.

These examples show that cat., F and JH are three different indexing systems. Their correspondences must be checked across the catalogue entry, the Concordance and the local catalogue. No number should be joined to another merely because it looks similar.

An interval, not an oracle

The result of dating research is not to hide uncertainty, but to show which clues already support one another and which still need to remain open. For the Paris period, letters give the moving track of a life, seasons and the city give the scene, canvas and pigments give material time, and brushwork gives continuity of work.

When these clues point to the same interval, we have not reached absolute certainty, but a judgement that can be checked. That judgement is more reliable than a beautiful precise date with no evidence behind it.