Histoire et mesure

Vol. XVII - N° 1/2, Varia

Usages et principes
Version imprimable   Signaler cette page  
David W. Galenson

Measuring Masters and Masterpieces

French Rankings of French Painters and Paintings from Realism to Surrealism

Notes de la rédaction

Note de l’auteur : I thank Martin Bruegel, Robert Jensen, Pierre-Michel Menger, Gerardo della Paolera, and participants in a seminar at EHESS for discussions, David Hwang for research assistance, and NSF for financial support. Work on this paper was begun when I was a visiting researcher at the Institut National de Recherche Agronomique, and I thank INRA for its hospitality.


Résumé

Mesurer la création artistique. La hiérarchie des peintres français et de leurs œuvres du Réalisme au Surréalisme.

Cet article prend en compte des illustrations reproduites dans des ouvrages français comme éléments de mesure de l'importance des peintures et aussi des peintres, à partir de l'exemple de 35 artistes majeurs qui travaillèrent en France au cours d'un siècle d'art moderne. Les classements ressemblent étroitement à ceux qui avaient été obtenus précédemment à partir d'une analyse identique conduite avec des ouvrages américains. Ils posent également une énigme : pourquoi quelques-uns des plus grands artistes n'ont-ils pas produit les peintures les plus célèbres, alors que d'autres peintres de renommée relativement mineure ont laissé des œuvres renommées ? La réponse semble résider dans une différence d'approche entre les artistes expérimentalistes, qui innovent continuellement, et les créateurs conceptuels qui réalisent des œuvres novatrices. Ce texte démontre également le bien-fondé de la quantification du succès artistique qui permet d'améliorer notre compréhension des origines de la créativité humaine.

Abstract

For 35 leading painters who worked in France during the first century of modern art, this paper uses illustrations in French textbooks as the basis for measuring the importance of both painters and individual paintings. The rankings closely resemble those obtained earlier from a similar analysis of American textbooks. They also pose a puzzle: why do some of the greatest artists not produce famous paintings, while some relatively minor artists produce famous individual works? The answer appears to lie in a difference in approach between experimental artists, who innovate incrementally, and conceptual innovators, who produce individual breakthrough works. This paper further demonstrates the value of quantifying artistic success, for doing so can improve our understanding of the sources of human creativity.

Texte intégral

In 1962, art historian George Kubler remarked that « the modern professional humanist is an academic person who pretends to despise measurement because of its 'scientific' nature, » and although the passage of four decades has now produced exceptions to Kubler's generalization in some disciplines within the humanities, art history is not prominent among them 1. In 1998, for example, curator Robert Storr of New York's Museum of Modern Art could declare that an artist's success « is completely unquantifiable » 2.

Yet Storr's belief is mistaken. As in other disciplines, quantitative evidence in art history has proven useful not only for constructing explanations, but for identifying problems to be explained. One recent study, for example, not only quantified artistic success, but helped to reveal an underlying structure of the creative process that deepens our understanding of the achievements of individual painters. The present paper extends this investigation, testing its conclusions by extending its analysis to a new body of data.

The earlier study used a quantitative analysis of the illustrations contained in published surveys of modern art to produce measurements of art historians' own judgments of the relative importance of modern French painters, and paintings 3. That study drew its evidence from 33 books published in English during the past three decades. One further question raised by this procedure concerned the possible impact of culture: would French scholars define the canon of French modern art differently from American and English scholars? The present study will answer this question by carrying out a parallel analysis based on surveys of modern art published in French.

1. Quantifying Artistic Success

The artists examined will be those considered by the earlier study. They are listed in Table 1. The sample was designed to include the most important modern painters who lived and worked in France for most or all of their careers, who were born between 1819 (the birth year of Gustave Courbet) and 1900. This was done by using five leading American textbooks on the history of modern art 4. The first step was to list all artists who had at least one painting reproduced in three or more of these five books. The 27 artists included in the list who had been born in France between 1819 and 1900 were placed in the sample. Another eight artists on the list who were born elsewhere during the same period but who spent substantial portions of their careers in France were also placed in the sample. In addition to Courbet and Manet, key figures in the transition to modern painting, the sample includes the central members of the groups - the Impressionists, the Post-Impressionists, the Fauves, the Cubists, and the Surrealists - that dominated modern painting France in the nineteenth century and well into the twentieth.

The evidence for the present study was drawn from all available books, published in French since 1963, that provide illustrated surveys of modern painting. A total of 31 such surveys were found; these are listed in the Appendix. The data set for this study was created by listing every reproduction of every painting shown in these books by all of the 35 artists in the sample.

There are fewer college texts published in French than in English, but there also appears to be a larger range of French reference works and surveys aimed at a general audience, and these provide a strong basis for this study. As in the earlier study, published surveys of art history were chosen as the source of the data in order to draw on the judgments of art scholars as to the most important painters and paintings. The dozens of authors represented include many distinguished academic art historians, critics, and curators of the recent past and present. Yet regardless of the distinction of the authors, all have taken the time and trouble necessary to communicate their views of the history of modern art, and the number of books is sufficiently large that no significant result depends on the opinions of any single author, or the emphasis of any one book.

Table 1. Artists included in this study

Artist

Country of birth

Year of birth

Year of death

Arp, Jean

France

1886

1966

Bonnard, Pierre

France

1867

1947

Braque, Georges

France

1882

1963

Cézanne, Paul

France

1839

1906

Chagall, Marc

Russia

1887

1985

Courbet, Gustave

France

1819

1877

Degas, Edgar

France

1834

1917

Delaunay, Robert

France

1885

1941

Derain, André

France

1880

1954

Duchamp, Marcel

France

1887

1968

Dufy, Raoul

France

1877

1953

Gauguin, Paul

France

1848

1903

Gogh, Vincent van

Holland

1853

1890

Gris, Juan

Spain

1887

1927

Léger, Fernand

France

1881

1955

Manet, Edouard

France

1832

1883

Masson, André

France

1896

1987

Matisse, Henri

France

1869

1954

Miró, Joan

Spain

1893

1983

Modigliani, Amedeo

Italy

1884

1920

Monet, Claude

France

1840

1926

Picabia, Francis

France

1879

1953

Picasso, Pablo

Spain

1881

1973

Pissarro, Camille

West Indies

1830

1903

Redon, Odilon

France

1840

1916

Renoir, Pierre-Auguste

France

1841

1919

Rouault, Georges

France

1871

1958

Rousseau, Henri

France

1844

1910

Seurat, Georges

France

1859

1891

Soutine, Chaim

Lithuania

1893

1943

Tanguy, Yves

France

1900

1955

Toulouse-Lautrec, Henri de

France

1864

1901

Vlaminck, Maurice de

France

1876

1958

Vuillard, Edouard

France

1868

1940

Whistler, James

United States

1834

1903

This investigation is analogous to a citation study. Yet using illustrations as the unit of observation has an advantage, in that illustrations are considerably more costly than written references. In addition to the greater cost of printing, authors or publishers must also bear the cost of obtaining permission to reproduce each painting, and a suitable photograph. This higher cost in both time and money tends to make authors more selective in their use of illustrations, and this may make illustrations an even more accurate indication than written references of what an author believes to be genuinely important. Thus for example psychologist Dean Keith Simonton's listing of the potential problems with the use of citations to measure scholarly importance includes the following: methodological articles tend to be cited more than other publications; research papers may cease to be cited when their results become incorporated into a discipline's common knowledge; papers originating in more prestigious institutions may be cited more often than articles of equal quality produced elsewhere; and scholars may gain high citation counts by writing papers on fashionable subjects 5. The paintings illustrated in textbooks will not be subject to analogous problems, because in effect they have been screened by the critical judgment of the scholars who are reviewing the history of art. These scholars will generally be able to distinguish major contributions from useful but relatively minor advances, and they will typically have the perspective to separate genuinely important and enduring accomplishments from those that have proved to be ephemeral and less consequential.

It might be objected that illustrations are chosen not according to the importance of the work, but instead to the ease of obtaining the relevant permissions and photographs. Yet for major artists, whose work has had many decades to find its way into public museums, the constraint posed by problems of access is not great. So for example 23 different works by Cézanne from just the collection of the Musée d'Orsay are illustrated in the 31 books used by this study, as are 24 works by Monet, and 23 by Degas; 16 different illustrated Matisses are held by the Musée National d'Art Moderne, and 25 illustrated Picassos are held in Paris by either the Musée National d'Art Moderne or the Musée Picasso. Scores of other paintings by these artists are of course held by other great museums 6. And the works owned by museums tend to be important ones, because curators – particularly at major museums – have little interest in acquiring unimportant works. Thus it seems clear that authors can readily choose among large numbers of important works in illustrating their books. Their decisions as to which to use will consequently tend to give a good indication of which they consider most important.

The use of published surveys of art history in the manner described above is aimed at identifying the central core, or canon, of early modern art. In art, as in science, the importance of a practitioner, or of a work, is a function of influence. Yet unlike scientific papers, paintings do not explicitly identify their influences. Identifying influences is consequently one of the key tasks of art historians. This may be difficult in some cases, and accordingly there can be differences of opinion among these scholars. Yet as the critic Clement Greenberg observed, in each generation there is a consensus in the judgment of quality in art: « The best taste is that of the people who, in each generation, spend the most time and trouble on art » 7. The analysis of large numbers of published surveys allows us to draw systematically on the opinions of scores of just such people, and thus effectively to poll large numbers of art experts on the issue of which painters, and paintings, they consider most important in the narrative of art history. Doing this proves to be an interesting way to begin to remedy the neglect of quantitative methods by the discipline of art history.

2. Cross-Cultural Canon Comparison

The two most basic rankings defining the canon of French modern art are straightforward: Table 2 ranks the artists by total illustrations, and Table 3 ranks individual paintings by the same measure. Each table also shows the corresponding American ranking of the respective painters and paintings.

The similarity between the French and American rankings is striking. Considering first Table 2, Picasso holds the top place in both France and the US, in both cases by a large margin over his closest competitors, who are in both cases Matisse and Cézanne, respectively. Although there are minor differences in their ordering, the top 10 painters are identical in both rankings. Larger discrepancies in order appear lower in the rankings, as the distances among painters in numbers of illustrations become smaller, but the rankings remain similar; so for example only one of the top 19 painters in the French ranking fails to appear among the top 20 in the American ranking.

Table 2. Ranking of artists by total illustrations

French rank

Artist

French illustrations

US rank

US illustrations

1

Picasso

206

1

335

2

Matisse

121

2

169

3

Cézanne

120

3

136

4

Monet

108

5

125

5

van Gogh

101

6(t)

116

6

Manet

97

4

130

7(t)

Braque

85

6(t)

116

7(t)

Gauguin

85

8

97

9

Renoir

75

10

74

10

Degas

74

9

81

11

Courbet

73

12

68

12

Bonnard

62

20(t)

29

13

Léger

61

15

51

14

Toulouse-Lautrec

58

16

40

15

Seurat

53

14

60

16

Delaunay

51

20(t)

29

17

Miró

49

13

64

18

Gris

41

28

19

19

Duchamp

40

11

72

20(t)

Chagall

37

18

35

20(t)

Derain

37

24

27

22

Rousseau

36

23

29

23(t)

Picabia

35

27

20

23(t)

Pissarro

35

17

38

25

Modigliani

34

34

10

26

Rouault

27

19

31

27(t)

Masson

26

33

12

27(t)

Redon

26

26

23

29(t)

Dufy

25

35

7

29(t)

Vuillard

25

29(t)

18

31

Whistler

23

20(t)

29

32

Soutine

22

31

14

33

Vlaminck

20

30

15

34

Tanguy

18

32

13

35

Arp

14

25

26

Source : This and subsequent tables are based on the data set constructed for this study. See text and appendix for description. The last two columns are from D. Galenson, 2002, p. 7.

Table 3 shows that French and American scholars are also in substantial agreement on the greatest individual paintings. Picasso's Les Demoiselles d'Avignon tops both rankings by a convincing margin, and four of the next five works in both rankings are the same, as French and American scholars agree on the great importance of Picasso's Guernica, Seurat's Grande Jatte, Manet's Déjeuner sur l'herbe, and Duchamp's Nude Descending a Staircase. Eight of the top 12 works in the French rankings appear among the highest American 11, and two more of those top French 12 are ranked in the top American 21. Only two of the 12 paintings most highly esteemed by French scholars – Courbet's Young Women and Rousseau's Snake Charmer – fail to appear among the 21 most favored by American scholars.

Table 3. Ranking of painting by total illustrations

French rank

French illustrations

Artist, Title

Date

Location

US rank

US illustrations

1

25

Picasso, Les Demoiselles d’Avignon

1907

New York

1

30

2(t)

18

Manet, Olympia

1863

Paris

8(t)

15

2(t)

18

Picasso, Guernica

1937

Madrid

2

25

4

17

Seurat, Sunday Afternoon on the Island of the Grande Jatte

1886

Chicago

3

24

5

16

Manet, Le Déjeuner sur l’Herbe

1863

Paris

4(t)

21

6

14

Duchamp, Nude Descending a Staircase, No. 2

1912

Philadelphia

4(t)

21

7

13

Monet, Impression, Sunrise

1872

Paris

19(t)

11

8(t)

12

Courbet, Burial at Ornans

1850

Paris

14(t)

12

8(t)

12

Duchamp, The Bride Stripped Bare by Her Bachelors, Even

1923

Philadelphia

7

16

10(t)

11

Courbet, Young Women on the Banks of the Seine

1856

Paris

--

--

10(t)

11

Gauguin, The Vision After the Sermon

1888

Edinburgh