Eighty-three, or Ninety-one
Bruegel painted an inventory in 1560 and nobody has agreed on its length since. The instability is not sloppiness, it is what happens when you catalogue without defining the unit.
Date
Category
Writer
The Kunsthistorisches Museum in Vienna owns a painting that contains, by its own account, eighty-three games. The Web Gallery of Art says about eighty-four. The German Wikipedia says ninety-one. Sandra Hindman, writing in The Art Bulletin in 1981, says more than ninety. Jeannette Hills, whose 1957 monograph is the foundational study, worked through seventy-eight. The English Wikipedia manages to say both "over 90" and "some eighty" on the same page, each with a citation.
The painting is Pieter Bruegel the Elder's Children's Games, finished in 1560. It has been in public collections and under scholarly attention for four centuries. Nobody has agreed on how long it is.
I want to argue that this is not a failure of counting. It is what an inventory does when the unit was never defined, and it is the most useful thing in the picture for anyone who builds systems that hold things.
Pieter Bruegel the Elder, Children's Games, 1560. Oil on oak panel, 116.4 × 160.3 cm. Kunsthistorisches Museum Vienna, Gemäldegalerie 1017. Public domain, via Wikimedia Commons.
What the object actually is
Before the argument, the facts, because most writing about this painting inherits its numbers from other writing about this painting.
Oil on oak panel. The museum gives two sets of dimensions: 116.4 × 160.3 × 1.2 cm for the measured panel, and 118 × 161 cm as the catalogue figure. Both are the museum's own. The 118 × 161 that saturates the secondary literature is not an error, it is the rounded number. Signed BRVEGEL 1560 at lower right, on the beam. Inventory number 1017 in the Gemäldegalerie, Room X. Acquired by Archduke Ernst in Brussels in 1594. The first textual mention is Karel van Mander's Schilder-boeck of 1604, which describes it simply as being about all sorts of children's games.
On the population, the museum is unambiguous and everyone else is not. The KHM says "over 230 children." The German Wikipedia offers 168 boys and 78 girls, which comes to 246 and carries no scholarly citation at all. The Web Gallery of Art says more than 250. Amy Orrock, whose 2012 article in the Journal of Historians of Netherlandish Art is the best recent treatment, says more than two hundred and leaves it there.
Why the counts disagree
The disagreement is not arithmetic. Nobody has miscounted. The counters are counting different kinds of thing, and the difference survives translation while the category does not.
Count | Source | What is being counted |
|---|---|---|
78 | Jeannette Hills, 1957 | Einzelinterpretationen: individual catalogue entries she chose to write |
80 | English Wikipedia's table | numbered rows, with sub-entries 14b, 17b, 39b to make things fit |
83 | Kunsthistorisches Museum | "different games" |
84 | Web Gallery of Art | "about 84 games" |
91 | Roberts-Jones, 1997 | Spielszenen: play scenes |
91 | Victor de Meyere, 1941 | spelletjes en bezigheden: games and activities |
more than 90 | Sandra Hindman, 1981 | "more than ninety actual games" |
Table 1. Seven counts of one painting. The column that matters is the third one.
Look at the two ninety-ones. Roberts-Jones counted scenes. De Meyere counted games and activities, and the Dutch en bezigheden is an open admission that the boundary would not hold. German Wikipedia then reports both as Kinderspiele, children's games. The number came through intact and the definition was dropped on the way.
Hindman's phrasing is the other tell. She writes "more than ninety actual games," and that adjective is load-bearing: she is filtering real games out of a larger pool of motifs, which means she already knew the pool and the filter were separate things. Her Appendix I is, in her own description, a concordance of competing identifications. The foundational scholarship was reconciling rival catalogues in 1981. The instability is not a modern internet artefact, it is original to the study of the painting.
Hills is the cleanest case of all, and the most misread. She explains her method plainly: she numbered the games in the picture, starting with the one in the lower left corner as number one. That is a reading device, a way of letting a reader find the thing she is describing. Later writers read the highest number in her list as a census. A numbered list is not a population figure, and the difference between those two readings has been quietly load-bearing in the literature for sixty years.
There is also the simple question of whether an activity is a game. Bathing in the river. Climbing a tree. A mock bridal procession. Riding a hobby-horse. Orrock notes the panel contains games played across the whole year, which is a heterogeneous body of material that resists any single taxonomy. You cannot count until you have decided, and the decision is not in the picture.
One count, no schema
This is the part I keep coming back to, because I have watched it happen in rooms with laptops in them.
Ask how many components are in a design system and you will get a number. Ask two people and you will get two numbers, and both will be right, because one is counting exported React components and the other is counting entries in the Figma library, and a dropdown is one thing in one place and four in the other. Ask how many pages a site has and the answer depends on whether a paginated archive is one page or thirty. Ask how many features shipped this quarter. Ask how many bugs are open.
In every one of those cases the argument presents as a disagreement about the total, and is actually a disagreement about the unit. The number is the visible part. The schema is the part nobody wrote down.
An inventory without a defined unit has no length. It has as many lengths as it has readers.
Bruegel is the four-hundred-year version of this. The painting is a catalogue that never shipped a schema, and so it has been recounted every generation, always honestly, never the same way twice.
The composition is built for the inventory, and it costs something
The museum is unusually direct about why the viewpoint is so high. Its own object text says that the bird's eye perspective "was the only way Bruegel could render visible the impressive crowd of figures." Not a symbolic choice. A display problem, solved.
Everything else follows from the same requirement. Orrock's description of the result is exact:
Despite these compositional structures there is no sense of narrative order to Bruegel's collection. The game motifs are all of a similar size and events at the center of the picture appear no more charged with importance than those at its edges.
Uniform scale. No focal point. Each group of players isolated from every other, with no excluded children, no idle spectators and no supervising adult to organise the reading. It is a grid of non-interacting cells wearing the costume of a scene.
That produces a specific and familiar failure, and Rose-Marie and Rainer Hagen named it precisely:
The picture offers no ideal standpoint. To keep an overview the viewer must step back from the picture. But the details are only legible from close up.
Overview and detail are mutually exclusive, and no amount of walking solves it. Anyone who has built a dashboard, a map, a large table or a zoomable canvas has met this exact bind, and has probably tried to solve it with a zoom control, which does not solve it. Zoom lets you choose which of the two you lose. It does not give you both.
The real solution is the one Bruegel did not use, because the technology was a panel of oak: a change in what is drawn at each scale, not only how large it is drawn. Semantic zoom rather than optical zoom. At a distance you should be seeing a different representation, not the same representation smaller.
The method generalises, and so does the problem
Pieter Bruegel the Elder, Netherlandish Proverbs, 1559. Oil on oak panel, 117.2 × 163.8 cm. Staatliche Museen zu Berlin, Ident. Nr. 1720. Public domain, via Wikimedia Commons.
A year earlier Bruegel applied the same method to language instead of play. Netherlandish Proverbs, 1559, now in Berlin, same format, same oak, almost the same size. And the count behaves exactly the same way: the Staatliche Museen say more than a hundred proverbs, the English Wikipedia says approximately 126.
The Wikipedia article even concedes the mechanism, noting that Bruegel may have included others which can no longer be determined because the language has changed. The unit of a proverb turns out to be no more stable than the unit of a game. Change the method's subject matter and the same instability reappears, which is decent evidence that it belongs to the method and not to the subject.
Dutch has a word for this class of picture: wemelbeelden, teeming pictures. It is a genre, and the genre's defining problem is enumeration.
Even the reading is disputed
I have been calling the painting an inventory, and I should say clearly that this is contested, because the journal is not worth much if I present a live argument as a settled one.
The museum endorses the inventory reading in strong terms, calling the composition "without precedent or parallel in the fine arts" and describing it as a painted encyclopaedia, adding pointedly that it carries no moralising undertones. The Web Gallery of Art goes further, calling it a catalogue and an enumeration without parallel in the history of art.
Hindman rejects this directly. She refuses to read it as simply a visual encyclopedia of games, arguing that visual cues signal the importance of specific games and the relationship of one game to another, and she reads the painting as a statement about human folly running through every age of life, not only childhood.
Orrock in turn criticises that whole school for a method problem worth quoting, since it is a failure mode that design writing commits constantly:
These moralizing iconological readings have now become dominant in the historiography of Children's Games, despite the obvious methodological flaw in using seventeenth-century emblems to decode a sixteenth-century painting.
Reading a 1560 painting through emblem books published decades later is the art-historical version of explaining a decision by quoting a framework that did not exist when the decision was made. It produces confident, coherent, undated nonsense, and it is extremely persuasive to read.
The same discipline applies to the old idea that this was the first panel of an unfinished series on the ages of man. That proposal is real, it belongs to Erica Tietze-Conrat, and it has been rejected, for a reason Orrock states without decoration: the painting contains games played throughout the year, so it resists being read as one season, and no other works by Bruegel survive to support the notion of an Ages of Man series. It is a historiographical episode, not a live hypothesis. Christian von Mechel's Four Seasons reading and Charles de Tolnay's have gone the same way.
What I actually take from it
Three things, all of which have cost me time by being learned late.
Define the unit before you report the number. A count that travels without its definition will be re-derived by every reader, and they will each get a different answer and assume yours was wrong.
A numbered list is a reading device, not a census. Hills numbered the games so that a reader could find them. Sixty years of writers treated her highest number as a population. If your documentation numbers things for navigation, expect the last number to be quoted as a total.
Overview and detail do not coexist by scaling. They coexist by substitution. If the far view and the near view show the same marks at different sizes, you have built the 1560 solution, and you have inherited its bind.
And one more, which is less a lesson than a thing I like. The single most honest sentence in the whole literature is Hindman's adjective. Actual games. One word, doing the work of admitting that the category was decided by a person, and that the person was her.
Sources
Object record, dimensions, inscription, provenance and the museum's own reading: Kunsthistorisches Museum, object 321. Amy Orrock, "Homo ludens: Pieter Bruegel's Children's Games and the Humanist Educators," JHNA 4:2 (2012). Sandra Hindman, "Pieter Bruegel's Children's Games, Folly, and Chance," The Art Bulletin 63:3 (1981), 447–475. Jeannette Hills, Das Kinderspielbild von Pieter Bruegel d. Ä. (1560), Vienna 1957, 2nd edition 1997. Rose-Marie and Rainer Hagen, Pieter Bruegel d. Ä., Taschen 1999, p. 33. Philippe and Françoise Roberts-Jones, Pieter Bruegel der Ältere, Hirmer 1997, pp. 220–221. Netherlandish Proverbs object record: Staatliche Museen zu Berlin.
I could not find a source for the figure of 86 games, which circulates online. If you know one, I would like to see it.
I write about art in Korean on my Naver blog, 베이니월드, where the same subjects get a different treatment.
Related reading
Latest Articles.
© Moc Mallen Studio
Thoughts, ideas, and perspectives on design, simplicity, and creative process.
