EMPIRICAL STUDIES OF THE ARTS, Vol. 28(1) 37-72, 2010 ON THE FORMATION OF CANONS: THE DYNAMICS OF NARRATIVES IN ART HISTORY VICTOR GINSBURGH Universitě Libře de Bruxelles and Universitě catholique de Louvain SHEILA WEYERS Universitě catholique de Louvain ABSTRACT The article illustrates the formation over time of the late 20th century canons of two schools that dominated all other European schools in their time: Italian Renaissance and Flemish Realism. Since most artists were discussed some 400 years ago, by Vasari in the second edition of his Vite and by van Mander in his Schilder-boeck, narratives by art historians can be followed over a long period of time. To explore the dynamic process of canon formation, we collected data on the presence and the greatness of a large number of artists in narrative works written by important art scholars at time intervals of roughly 75 years, so that the 400 years elapsed between 1600 and 2000 are spanned as best as possible. At least half of the artists in the two canons were already there 400 years ago. There are several cases of wrong attributions or of new technical discoveries that prevented some names to be canonical any sooner. There are also artists who art historians learned to appreciate or to understand better, and who entered at much later times. Finally, the number of names that entered or were moved up in the canon because their works acquired new properties in the light of works by artists that followed them is not very large. This appears to be in contradiction with the frequent suggestion that canons are continuously moving and that no artist can survive forever. 37 © 2010, Baywood Publishing Co., Inc. doi: 10.2190/EM.28.1.d http://baywood.com 38 / GINSBURGH AND WEYERS In her paper on canon formation, Silvers (1991, p. 212) suggests that "understanding how evaluative critical judgment evolves might be supposed to require detailed empirical study more properly pursued in disciplines other than philosophy because the nature of the process which forms canons is sociological, political or economic, rather than autonomously aesthetic." In this article, we try to contribute by illustrating the formation of the late 20th century canons of two groups of artists (Italians and Flemish) who, according to Panofsky (1971) dominated all other European schools during the Renaissance. Since most of them were discussed some 400 years ago, by Vasari (1981/1568) in the second edition of his Vite and by van Mander (1604) in his Schilder-boeck, narratives by art historians can be followed over a long period of time. To explore the dynamic process of canon formation, we collected data on the presence and the greatness of a large number of artists in narrative works written by important art scholars at time intervals of roughly 75 years, so that the 400 years elapsed between 1600 and 2000 are spanned as best as possible. Though the choice of art scholars as "true judges" may be "embarrassing" (Hume, 1965/1757, p. 17), we shall consider them as our trustworthy (though fallible) indicators of the best artworks, or artists, whose identification emerges as a consequence of the passage of time and leads to canonical status. How Do Artists (or Works) Attain Canonical Status? Silvers (1991, pp. 212-213) describes several paths that may lead to canonical status. A work (or an artist) may qualify in one of the following three ways: (a) failing, despite systematic scrutiny, to reveal defects or disagreeableness sufficient to be disqualified; or (b) revealing previously unnoticed meritorious or agreeable properties sufficient to qualify; or (c) acquiring valuable properties sufficient to qualify it. She points out that (a) and (b) account for this process in terms of permanence or changes in the opinions of art scholars (traditionalism), while (c) accounts for events that took place after a work was produced and change its properties (revisionism). According to traditionalist art theorists, all properties are present when the work is conceived and realized, though their importance may have been overlooked. Leonardo da Vinci's oeuvre is an example that satisfies criterion (a). According to Grove's Dictionary (1996, vol. 19, p. 196) "there has never been a period in which Leonardo's greatness has not been acknowledged." A similar argument can be invoked when attributions are revised. Duccio appeared in the canon after Berenson gave him the Madonna di Ruccellai, previously thought by Cimabue. The Master of Flemalle has benefited from both a rediscovery and attributions of works thought by the young van der Weyden. In such cases, changes in canonical status are due to epistemic reasons. ON THE FORMATION OF CANONS / 39 Revisionists advocate that some properties or attributes may be added to existing works by newly created works, or that earlier works "acquire salience in relation to the proper understanding of [new works], which they naturally didn't have before" (Levinson, 1996, p. 268). Silvers (1991) illustrates what happened with Rubens' figures after Renoir and Picasso. Rubens himself considered his figures to be coarse, and this was accepted for centuries after he painted them. Today, writes Silvers, "when art's history embraces treatments of the human figure such as those painted by Renoir and Picasso, Rubens' treatments are transfigured to become fluently refined and elegantly vital . . . the composition of Rubens' paintings remain[s] the same, but the works' aesthetic attributes change, develop, transmogrify or evolve" (p. 217). Junod (1995) similarly suggests that Vermeer was rediscovered in the mid-19th century due to the closeness of his work to pre-impressionist sensitivity. Whether such situations change the properties of the work itself is hotly debated among art philosophers. Levinson (1990, p. 194), for example, argues against revisionism, and brings the revisionist argument to an extreme with another example: after Cubist painting came into existence, "the non-Cubist mode of depiction of, say, Holbein's Ambassadors suddenly appears as an artistically relevant attribute of this painting," though the painting was produced in 1533. Both traditionalism and revisionism are present in the narratives of art historians who, by continuously provoking our attention, are among the most important contributors to the formation of canons. That traditionalism is present is obvious, since historians evoke what occurred before and during the creation of the artworks that they describe. Revisionism comes into narratives because art historians also take into account what happened after the creation of the works. Discussing Goya's influence on Bacon in an article devoted to Bacon qualifies Bacon. But Bozal's (1997) observation in his monograph on Goya that the "horribly open devouring mouth in Saturn is a prelude to the howling mouths of Bacon," may be an addition, even if only second-order, to the fame of Goya. However, one can argue whether this changes Goya's Saturn itself, or whether it merely changes our vision of the work.1 Therefore, as we shall see, it is often difficult to decide whether a work qualifies by criterion (b) or (c). The above analysis of canon formation implies that artworks are endowed with properties, a view that is not shared by all art theorists. Bourdieu (1983, 1996; see also Hutter & Shusterman, 2006, p. 193) argues that evaluation, and thus value, is arbitrary, even if it stands the test of time, because it is based on motivations imposed by the social and political structures of the cultural hierarchy. It is objective but only as a social fact: the artistic field is contained within the field of power, which is itself situated within the field of class relations (Bourdieu, Note that, in some interviews, the Spanish art historian Junquera, author of a monograph on Goya's black paintings (Junquera, 2003) casts doubt about Goya being their author. 40 / GINSBURGH AND WEYERS 1983, p. 319). Accordingly, there are no criteria that allow determining the intrinsic quality of a work, but only professional judges who "possess the socially accepted authority to ascribe specific properties to a work... and how it should be ranked" (Van Rees, 1983, p. 398; and see Rajagopalan, 1997; Van Peer, 1996). METHOD To describe the process of canonization, we follow over time the presence of artists (measured by the length of the entry or the number of citations devoted to each of them; length of entries and citations are always transformed into ranks) in a certain number of art histories, encyclopedias, and art dictionaries. Such an approach had already been suggested by Teyssedre (1964, p. 187), the expert of the French art critic de Piles (1635-1709), who writes that "even if ratings are difficult to assess, one could just look at whether artists are quoted or not." Milo (1986) uses the length of entries in encyclopedias and dictionaries over time, to study whether 17th century French painters who are praised today were already so between 1650 and 1750. See also Verdaasdonk (1983, 2003), Simonton (1998), or Ginsburgh and Weyers (2006) who perform similar analyses for literary works, operas, or Italian Renaissance painters. Using such a method, without any appeal to the details of the narratives of art historians, may seem unusual, and it happens that the length of an entry devoted to an artist may be due to criticizing him instead of praising him. One such example is that of Felibien's lukewarm comments on both Veronese and Tintoretto, since they were both colorists, while he considered disegno to be superior. Felibien nevertheless devotes more space to Veronese than to Andrea del Sarto and Leonardo (two Florentine artists), and Tintoretto gets almost as much space as Leonardo. The opposite can also happen: Vasari had a very favorable opinion of Corregio, but does not rank him very high in terms of the length of his vita. However, we take comfort in the fact that writing takes time and effort, and it is likely that "professionals will not devote labor and attention, generation after generation, to sustaining [artworks] whose life-functions have terminated" (Coetzee, 2002, p. 18). The alternative is using hermeneutics. These make it difficult, and often subjective, to decide who is part of the canon. Consider, for example, the following description given in Giorgione's entry by Grove (1996, vol. 12, p. 677): Giorgione received "the highest praise from Vasari, who, although disapproving of his method of painting without drawing, regarded innovative softness and suggestiveness of his handling admirably natural . . . and placed his Life close to the beginning of Part III of the Vite, between those of the other great pioneers of maniera moderna, Leonardo and Corregio." Still, the entry devoted to Leonardo is 2.6 times longer than the one devoted to Giorgione, and Corregio's is much shorter than Giorgione's. The more quantitative approach that we take, produces clear-cut answers that also have the advantage of allowing comparisons over time. The ranking that we base on ON THE FORMATION OF CANONS / 41 entries or citations (2 for an artist ranked among the top 50; 1 for an artist who is described without being among the top 50; 0 for the one who is only mentioned or even ignored) is a way of representing what Westphal (1993, p. 436) describes as the degree of canonicity: Status as a classic is not an either/or matter, but a matter of degree. At any given time, the canon is represented by a series of concentric circles. At the center are texts with the highest degree of canonicity, while at the periphery are those whose classical status is most tenuous. This means that historical changes in the canon are not simply matter of inclusion and exclusion, but also matters of location between the center and the periphery. The Italian Renaissance and Flemish Realism Canons at the End of the 20th Century For the sake of illustration, and without going into the aesthetic versus sociopolitical debate of defining the canon, we assume that the late Twentieth Century canons are defined by the names that appear in Grove's (1996) entries for the Italian Renaissance (vol. 16, pp. 654-668) and for Flemish painting (vol. 3, pp. 551-562). This generates two lists of 125 and 129 names, respectively. In both lists reproduced in Tables 1 and 2, artists are ranked according to the length of their individual entries in Grove's Dictionary. Here are some examples that illustrate how the tables should be read. For Italian art, Michelangelo is the artist whose entry is the longest (Grove Rank 1), Leonardo and Giotto come next, and Paolo di Giovanni is ranked 125th. Rubens is first in the Flemish canon, followed by Breughel and Van Dyck, while Clara Peeters has rank 129. Further details are given in Appendix 1. Each list provides dates of birth and death, or the period during which the artist flourished. This is especially useful in the case of Flemish painters who were born too late to be known by van Mander. For Italians, we also give some information on the city or region in which they were active (Central Italy, Florence, Northern Italy, Siena, and Venice) since their different styles are known to have influenced Vasari's writings, who preferred Florentine to North Italian and Venetian painting. This approach is different from Cutting's (2006a, 2006b) for the following reasons: (a) Cutting's purpose is to construct a canon (for Impressionist painters), while we take it as given by Grove; and (b) he looks at different sources (number of images, presence in museums and collections, writings of scholars, etc.) to determine who and what was influential in defining the canon; we only look at narratives by selected scholars and, given the 400 years that are embraced in our article, we are rather concerned with the dynamics (when) and the whys of its formation. 42 / GINSBURGHAND WEYERS Table 1. The Italian Canon Grove Active Rank Name in Born Died 1 Michel Angelo (Buonarroti) F 1475 1564 2 Leonardo da Vinci F 1452 1519 3 Giotto di Bondone F 1267 1337 4 Raphael Sanzio C 1483 1520 5 Tiziano Vecellio V 1485 1576 6 Tintoretto (Robusti) Jacopo V 1519 1594 7 Mantegna Andrea N 1460 1506 8 Bellini Giovanni V 1431 1516 9 Veronese Paolo V 1528 1588 10 Giorgione (Ziorzi) da Castelfranco V 1477 1510 11 Botticelli Sandro F 1445 1510 12 Bramante Donato N 1444 1514 13 Giulio Romano C 1499 1546 14 Masaccio Tommaso F 1401 1428 15 Angelico Fra Giovani F 1395 1455 16 Piero della Francesca C 1415 1492 17 Duccio du Buoninsegna S 1278 1319 18 Correggio (da) Allegri Antonio N 1260 1276 19 Francesco (Maurizio) di Giorgio Martini S 1439 1501 20 Verrocchio Andrea del F 1435 1488 21 Perugino (Vanucci) Pietro C 1450 1523 22 Martini Simone S 1284 1344 23 Bellini Jacopo V 1400 1470 24 Ghirlandaio Domenico F 1448 1494 25 Lippi Filippino F 1457 1504 26 Domenico Veneziano F 1405 1461 27 Salviati Francesco F 1510 1563 28 Sarto Andrea del F 1486 1530 29 Lotto Lorenzo V 1480 1556 30 Masolino da Panicale F 1383 1435 31 Lippi Filippo F 1406 1469 ON THE FORMATION OF CANONS / 43 Table 1. (Cont'd.) Grove Active Rank Name in Born Died 32 Uccello Paolo F 1397 1475 33 Sebastiano del Piombo V 1485 1547 34 Bartolommeo Fra della Porta F 1472 1517 35 Antonello da Messina C 1430 1479 36 Tura Cosimo N 1430 1495 37 Gentile da Frabriano N 1385 1427 38 Signorelli Luca C 1450 1523 39 Carpaccio Vittore V 1460 1525 40 Bassano Jacopo V 1510 1592 41 Bronzino Agnolo F 1503 1572 42 Lorenzo Monaco F 1370 1425 43 Roberti Ercole de N 1455 1496 44 Parmigianino Francesco N 1503 1540 45 Castagno Andrea del F 1419 1457 46 Cimabue Cenni di Pepo F 1240 1302 47 Paolo Veneziano V 1333 1362 48 Pisanello (Pisano) Antonio N 1395 1455 49 Pollai(u)olo Antonio F 1432 1498 50 Cavallini Pietro C 1240 1330 51 Sassetta Stefano di Giovanni S 1400 1450 52 Lorenzetti Ambrogio s 1317 1347 53 Bordone Paris V 1500 1571 54 Moretto (da Brescia) Alessandro N 1498 1554 55 Gozzoli Benozzo F 1420 1497 56 Lorenzetti Pietro S 1306 1345 57 Pollai(u)olo Piero F 1441 1496 58 Cione (Orcagna) Andrea di F 1315 1368 59 Guido(ne) da Siena S 1262 1270 60 Pinturicchio (di Betto) Bernardo c 1452 1513 61 Piero di Cosimo F 1461 1521 62 Cossa Francesco del N 1435 1476 44 / GINSBURGHAND WEYERS Table 1. (Cont'd.) Grove Active Rank Name in Born Died 63 Beccafumi Domenico S 1484 1551 64 Pontormo Carrucci Jacopo da F 1441 1496 65 Rosso Fiorentino Giovanni F 1494 1540 66 Dossi Dosso N 1490 1542 67 Gaddi Taddeo F 1320 1366 68 Bellini Gentile V 1429 1507 69 Coppo di Marcovaldo F 1260 1276 70 Crivelli Carlo V 1430 1495 71 Vecchietta Lorenzo di Pietro di Giovanni s 1410 1480 72 Daddi Bernardino F 1320 1348 73 Foppa Vincenzo N 1427 1515 74 Tito Santi di F 1536 1602 75 Abate (Abatti) Niccolo dell N 1509 1571 76 Francia Francesco N 1450 1517 77 Zuccaro Federigo C 1540 1609 78 Melozzo da Forli C 1438 1494 79 Menabuoi Giusto de N 1349 1390 80 Moroni Giovanni Battista N 1520 1574 81 Pordenone Giovanni Antonio V 1483 1539 82 Vivarin Antonio V 1440 1484 83 Baldovinetti Alesso F 1425 1499 84 Pesellino Francesco di Stefano F 1425 1457 85 Gaddi Agnolo F 1369 1396 86 Jacobello del Fiore V 1400 1439 87 Schiavone (Meldolla) Andrea V 1433 1504 88 Altichiero di Domenico da Zevio N 1369 1393 89 Master of the St Francis Legend C fl 1290 90 Pitati Bonifazio dei V 1487 1553 91 Vivarini Alvise V 1442 1503 92 Rosselli Cosimo F 1439 1507 93 Taddeo di Bartolo s 1362 1422 94 Muziano Girolamo c 1532 1592 ON THE FORMATION OF CANONS / 45 Table 1. (Cont'd.) Grove Rank Name Active in Born Died 95 Colantonio Niccolo C 1420 1460 96 Matteo di Giovani S 1430 1495 97 Maso di Banco F 1335 1350 98 Allori Alessandro F 1535 1607 99 Torriti Jacopo C 1270 1300 100 Domenico di Bartolo (Ghezzi) S 1400 1445 101 Daniele (Ricciarelli) da Volterra c 1509 1566 102 Sano di Pietro s 1405 1481 103 Andrea da Firenze F 1346 1379 104 Landi (del Pogio) Neroccio s 1447 1500 105 Berlinghieri father and son c 1228 1274 106 Apollonio di Giovanni F 1416 1465 107 Angu(i)ssola Sofonisba N 1532 1625 108 Lorenza Veneziano V 1356 1379 109 Giunta Pisano C 1236 1254 110 Barna (Berna) da Sienna S 1330 1350 111 Rusuti Filipo C 1297 1317 112 Pulzone Scipione C 1544 1598 113 Squarcione Francesco N 1395 1468 114 Cione Nardo F 1320 1365 115 Master of 1419 F fl 1419 -30 116 Scheggia Giovanni F 1406 1480 117 Be(o)rgognone Ambrogio N 1453 1523 118 Santi Giovanni N 1435 1494 119 Pino Paolo F fl 1534 -65 120 Zelotti Battista N 1526 1578 121 Avanzi Jacopo N 1363 1384 122 Cavalori Mirabello d'Antonio F 1535 1572 123 Macchietti (del Crucifissaio) Girolam F 1535 1592 124 Erri Agnolo + Bartolommeo N 1442 1497 125 Giovanni Paolo di S 1345 1441 Notes: C: Center; F: Florence; N: North; S: Sienna; V: Venice. See also text, fl = floruit, 46 / GINSBURGHAND WEYERS Table 2. The Flemish Canon Grove Rank Name Born Died 1 Rubens Pieter 1577 1640 2 Breughel Pieter 1525-30 1569 3 Dyck Anthony van 1599 1641 4 Weyden Rogier van der 1400 1464 5 Eyck Jan Van 1381 ? 6 Bosch Hieronymus 1450 1516 7 Goes Hugo van der 1440 1482 8 Gossart (Mabuse) Jan 1478 1532 9 Master of Flemalle fl 1420 -40 10 Jordaens Jacob 1593 1678 11 Memling Hans 1435 ? 12 Bouts Dieric I 1415 1475 13 Metsys Quinten 1466 1530 14 Orley (Von Brüssel) Bernard 1488 1541 15 David Gerard 1460 1523 16 Vos Maarten De 1532 1603 17 Justus van Gent (Wassenhove) fl 1420 -40 18 Janssen Abraham 1575 1632 19 Mor van Dashorst Antonis 1516 1576 20 Snyders Frans 1579 1657 21 Brouwer Adriaen 1605 1638 22 Cleve Joos Zotte ? 1540-1 23 Patinir Joachim 1480 1524 24 Teniers David the Younger 1610 1690 25 Thulden Theodoor 1606 1676 26 Hemessen Jan Sanders 1519 1556 27 Vos Cornells de 1584 1651 28 Vredeman de Vries Hans 1527 1606 29 Vellen (Velart) Dirk 1480 1547 30 Breughel Jan Velvet I 1568 1625 31 Floris Frans 1519 1570 32 Quellinus Erasmus 1607 1678 33 Franken Frans II 1581 1642 ON THE FORMATION OF CANONS / 47 Table 2. (Cont'd.) Grove Rank Name Born Died 34 Breughel Pieter II 1564-5 1637-8 35 Christus Petrus 1410 1475-6 36 Vos Paul de 1591 1678 37 Isenbrandt Adriaen ? 1551 38 Crayer Gaspar 1584 1669 39 Veen Otto 1556 1629 40 Massys (Metsys) Jan 1509 1575 41 Benson Ambrosius 1500 1550 42 Aertsen Peter 1507-8 1575 43 Fyt Jan 1611 1661 44 Diepenbeeck Abraham 1596 1675 45 Eyck Hubert Van 1385-90 1426 46 Heem Jan Davidsz 1606 1683 47 Valckenborch Lucas I 1535 1598 48 Coecke van Aelst Pieter 1502 1550 49 Mostaert Jan 1475 1555 50 Coninxloo Gillis 1544 1604 51 Uden Lucas 1595 1672 52 Hoecke Jan 1611 1651 53 Stockt Vrancken Van 1420 1495 54 Coter Colijn 1480 1520 55 Beuckelaer Joachim 1534 1574 56 Weyden Goswijn Van der 1465 1538 57 Momper Josse 1564 1635 58 Coques Gonzales 1618 1684 59 Coxcie Michiel 1499 1592 60 Blondeel Lancelot 1488 1581 61 Bles Herri met de 1480 1550 62 Vermeyen Jan 1500 1559 63 Wildens Jan 1584 1653 64 Peeters Bonaventura I 1614 1652 65 Pourbus Pieter 1523 1584 66 Provost (of Mons) Jan 1465 1529 48 / GINSBURGHAND WEYERS Table 2. (Cont'd.) Grove Rank Name Born Died 67 Lombard Lambert 1505 1566 68 Baien Hendrik I 1575 1632 69 Rombouts Theodoor 1597 1637 70 Massys (Metsys) Cornells 1511 1557 71 Boeckhorst Jan 1604 1668 72 Rijckaert David III 1612 1661 73 Loon Theodoor 1581 1667 74 Cock Jan Wellens 1490 1527 75 Siberechts Jan 1627 1703 76 Master of Legend of St Lucy ? ? 77 Craesbeeck Joos 1606 1654-60 78 Seghers Daniel 1590 1661 79 Master of Frankfurt 1460 1533 80 Bouts Albrecht 1452 1549 81 Master of Legend of St Mary Magdalena ? ? 82 Schut Cornelis I 1597 1655 83 Frémal Bertholet 1604 1675 84 Gassel Lucas 1495-00 1570 85 Key Willem Adriaens 1515 1568 86 Backer Jacob 1555 1585 87 Oost Jacob I 1603 1671 88 Cleve Maarten 1527 1581 89 Reymerswaele Marinus 1490 1567 90 Dalem Cornelis van 1530 1573 91 Grimmer Jacob 1526 1590 92 Master of the Legend of St. Barbara fl 1420 -40 93 Master of the Legend of St Catherine fl 1420 -40 94 Mostaert Gillis 1528 1598 95 Pepyn Maarten 1575 1642 96 Sellaer Vincent 1538 ? 97 Grimmer Abel 1570 1618-9 98 Valckenborch Maarten I 1534 1612 99 Beert Osias 1580 1624 ON THE FORMATION OF CANONS / 49 Table 2. (Cont'd.) Grove Rank Name Born Died 100 Boudolf Jan (Jean de Bruges) 1368 1381 101 Master of the Joseph Sequence fl 1420-40 102 Bol Hans 1534 1593 103 Sallaert Anthonis 1580 1650 104 Steenwijk Hendrik Van I 1550 1603 105 Coene Jacques fl 1420-40 106 d'Arthois Jacques 1613 1686 107 Es Fopsen Van 1596 1666 108 Francken Ambrosius 1 1544 1618 109 Mandijn Jan 1500 1559 110 Frankcen Frans I 1542 1616 111 Del Monte Deodaat 1582 1644 112 Master of 1499 1499 ? 113 Master of the Morisson Triptych 1525 ? 114 Thielen Jan Philips 1618 1667 115 Master of the Legend of St Ursula fl 1420-40 116 Thys Pieter 1624 1677 117 Huys (Hus) Pieter 1520 1584 118 Master of the Baroncelli Portraits 1489 9 119 Heuvel Antoon Van den 1600 1677 120 Veerendael Nicolaes 1640 1691 121 Minderhout Hendrik 1632 1696 122 Janssens Jan 1590 1650 123 Liemacker Nicolaas 1601 1646 124 Eertbeld Andries Van 1590 1652 125 Utrecht Adriaen Van 1599 1652 126 Luyckx Frans 1604 1668 127 Eyck Lambert Van fl 1420-40 128 Neeffs Pieter I 1578 1656 129 Peeters Clara 1589 1657 50 / GINSBURGH AND WEYERS On Narratives Scholarship guided our list of art historians whom we consider as our "true judges." They were also chosen for their overall coverage of Italian or Flemish painting (which led us to discard historians who deal with local or regional issues, as well as monographic works). For Italy, the 20th century canon is formed by painters whose names appear in the entry on Italian painting in Grove (1996, vol. 16, pp. 654-668). For each of them we collected the entries (or number of citations) in Vasari's Vite (1981/1568), Felibien's Entretiens (1967/1725), Lanzi's art history of the Quattrocento (1824/1789), Burckhardt's Cicerone (1855), Berenson (1926, 1894, 1896, 1897, 1907), Chastel (1995/1956), and Grove's Dictionary itself. Vasari's Vite appear an undisputable choice, though, as has often been pointed out, the work is biased toward Florentine artists, and gives less credit to Venetians, like Giovanni Bellini or Veronese. Felibien (1967/1725, p. 40), who writes 100 years after Vasari, is considered the father of art history and art criticism in France. He mentions that "as far as modern painters are concerned, I merely follow what Vasari, Borghini, Ridolfi, the cavaliere Baglione and a few others have amply described, and with whom I agree," though he discusses at great length Venetians. Under the influence of Winckelmann, Lanzi makes art history into a discipline that does more than describing the lives of artists. His work encompasses Italy as a whole. He classifies artists according to schools, including a large number of local schools, and tries to convey an impartial view of history, putting aside his personal neo-classical tastes. According to Bazin (1986, p. 91), Lanzi's work contains 3,000 names of artists. Lanzi is proud to claim that he makes no selection but also discusses mediocre artists, who, given their relations with the "great," do also participate. Burckhardt is almost unanimously considered to be one of the greatest historians of the Italian Renaissance, and is representative of Kulturgeschichte, a movement, which suggests that art produced in an era cannot be separated from the society by and in which it is produced. He was also deeply influenced by Vasari. In Kultur der Renaissance in Italien, Burckhardt mentions having copied over 700 excerpts from Vasari's Vite, and inserted them at the right places in his own book (Gombrich, 1969). By choosing Berenson to represent the views of the early 20th century, we privilege connoisseurship. Berenson's choices were, nevertheless, very influential on the whole century. To represent the mid-20th century, we selected Chastel's celebrated work on the Italian Renaissance. Chastel is considered the most influential art historian in France after World War II. We end with Grove's encyclopedic Dictionary narratives. We followed a similar procedure for Flemish painting, picking all the painters mentioned in the entry on Flemish painting in Grove (1996, vol. 3, pp. 551-562). For each artist, we retrieved the length of the entry in van Mander (1604), Sandrart (1675), Descamps (1753, 1760-1764), Fiorillo (1815), Immerzeel (1842, 1855), ON THE FORMATION OF CANONS / 51 Wurzbach (1906-1911), van Puyvelde (1953, 1962) van Puyvelde and van Puyvelde (1970), and Grove's Dictionary. Max Friedländer (1924-1937) could have been an obvious choice as representing the early 20th century, but his writings are essentially concerned with old Netherlandish painters and do not cover the second half of the 16th century. Though van Mander is considered as important as Vasari, he wrote his Schilder-boeck at a time where great Flemish painters such as Rubens (born in 1577), Jordaens (1593), or Van Dyck (1599) were very young and could hardly have been known by him. Therefore, not surprisingly, van Mander missed many great names of the Golden Age of Flemish painting who belong to the Flemish canon. The painter Sandrart, who represents the late 17th century, is famed for his biographical writings inspired by Vasari and van Mander. Descamps, a French painter and dealer, was also the very successful writer of La vie des peintres flamands, though, according to Grove (1996, vol. 8, p. 788), his work contains many inaccuracies. Fiorillo, a painter and professor of art history in Göttingen, was influenced by Lanzi, who emphasizes the "compilation of information at the expense of interpretive synthesis" (Grove, vol. 11, p. 118). Immerzeel, a Dutch art and books dealer, wrote his Levens on the basis of existing biographies, but also used unpublished manuscripts and documents. Wurzbach, an Austrian art historian, produced his Niederiändishes Künstler Lexikon, considered a standard dictionary of Flemish (and Dutch) artists. Van Puyvelde was chief curator of the Royal Museums of Fine Art of Belgium in Brussels from 1927 to 1948. His three volumes on Flemish painting represent the views of the mid-20th century, though Bazin (1986, p. 502) considers his judgments too personal and always taking a view opposite to generally accepted ideas. Grove's Dictionary closes the list of our narrators. In both cases, three art historians or art histories represent the 20th century, while the 200 or 300 years between Vasari (or van Mander) and Berenson (or Wurzbach) are spanned by four names only. This is due to two main reasons. First, we are obviously more interested in how our era evaluates art created during the Renaissance. Second, if, as suggested by Junod (2002), the past can be rediscovered through contemporary works, it may be useful to examine who and what has been rediscovered in the wake of the largest possible set of narratives, and have a finer grid for more recent years. Table 3 gives a summary view of the database that was set up. The Dynamics of Canonization We now concentrate on artists who are ranked 1 to 50 in Grove's Dictionary and try to understand how the two late 20th century canons were formed. For each artist, we compute the rank (based on length of the entry, or number of citations) given by previous art historians (Vasari to Chastel for Italians, and van Mander to van Puyvelde for Flemish painters). Each artists is assigned code "2" if he is 52 / GINSBURGHAND WEYERS Table 3. Overview of the Data Approximate Historian Life of Publication Form of period (author) historian date of book data used Italian artists 1550 Vasari 1511-1574 1550, 1568 Length of entry 1650 Felibien 1619-1695 1659-1689 Length of entry 1775 Lanzi 1732-1810 1789 Length of entry 1850 Burckhardt 1818-1897 1855 Citations 1900 Berenson 1865-1959 1894-1907 Citations 1950 Chastel 1912-1990 1956 Citations 2000 Grove 1996 Length of entry Canon Grove 7996 Cited in Grove article 'emish artists 1600 Van Mander 1548-1606 1604 Length of entry 1675 Sandrart 1606-1688 1675 Length of entry 1750 Descamps 1715-1791 1753 Length of entry 1800 Fiorillo 1748-1821 1815 Length of entry 1850 Immerzeel 1776-1841 1842, 1855 Length of entry 1900 Wurzbach 1845-1915 1911 Length of entry 1950 Van Puyvelde 1882-1965 1953-1970 Length of entry 2000 Grove 1996 Length of entry Canon Grove 7996 Cited in Grove article ranked among the top 50 by a given art historian. If he is discussed at sufficient length, but is not part of the top 50, his code is "1." If he is ignored or only very briefly mentioned (for example in passing in the life or entry devoted to another painter), his code is "0." Consider for example (Jacopo) Bassano in Table 4. Lanzi and Berenson rank him among the top 50, he therefore is assigned a "2." Felibien, Burckhardt, and Chastel discuss him, but none of them ranks him among the top 50 (code "1"). He has no vita or notice in Vasari's Vite, and is thus assigned code "0." Michelangelo is always among the top 50 (code "2"). Tables 4 (Italians) and 5 (Flemish) are also constructed in such a way that they make the dynamics more transparent. For Italians, we rank first those among the top 50 by Vasari, though some may disappear and reappear again later (Fra Angelico or Verrochio, for example). Then we move to those who were not ranked as such by Vasari, but introduced later by Felibien (starting with Tintoretto), and we keep going that way for Lanzi, Burckhardt, Berenson, and Chastel. The first column in the table (Grove Ranks) gives the order in which Grove ranks artists. Michelangelo, for example, is first, Mantegna is seventh, etc. ON THE FORMATION OF CANONS / 53 Table 5, devoted to Flemish painters, is constructed in the same way. Drawing mainly from Grove's (1996) "critical reception" or "posthumous reputations" remarks, we now turn to the analysis of the two canons using Silvers' criteria. The choice of 50 as the number of painters in Westphal's first circle of the canon may look arbitrary and large when compared to the seven names that Cutting considers as the first circle of the Impressionist canon. This choice is discussed in Appendix 2. Italian Painters To begin with, it is worth pointing out that Vasari devotes a vita to 45 out of the 50 canonic painters; two Venetian artists (Tintoretto and Giorgione) have notes that are however longer than many vite, the names of two others (Bassano and Tura) are at least mentioned in other lives or notes. Only one of them (Paolo Veneziano) is not referred to at all. We follow the stages set up in Table 4, in which artists enter the canon with Vasari (1981/1586), Felibien (1967/1725), Lanzi (1824/1789), Burckhardt (1855), Berenson (1926/1894, 1896, 1897, 1907), Chastel (1995/1956), and Grove (1996), at various moments in time. Vasari Eight artists from the Italian Renaissance (Michelangelo, Leonardo, Giotto, Raphael, Titian, Mantegna, Perugino, and Andrea del Sarto) are always among the top group of 50. These are canonized in terms of Silvers' (a) criterion. A second group of 17 artists (Botticelli, Bramante, Giulio Romano, Masaccio, etc.) are among the 50 first artists in Vasari's Vite, but disappear from the top group at some point after Vasari and reappear later. However, with the exception of Carpaccio and Bartolommeo della Porta (ignored by Felibien), Parmigianino (ignored by Burckhardt), and Salviati (ignored by Berenson), they are always present even if not among the top 50. These dynamics of being present, leaving, and reappearing in the top group are not surprising, and are evoked by many art historians. With the exception of some (Leonardo, Giotto), most artists had darker periods. The case of Botticelli has often been underlined, but is far from being unique. Even Michelangelo was subject to negative criticism. Ludovico Dolce (1508-1568) "unfavourably compared his narrow expertise in depicting the male nude with the greater variety displayed by Raphael and Titian" (Grove, 1996, vol. 21, p. 459); some 150 years later, the critic Milizia thought that he did not understand anatomy (Junod, 1995). Even Raphael was mistreated at some point, as the "normative status that Poussin and Ingres and their followers gave to his art has certainly done much to diminish his popularity" (Grove, 1996, vol. 25, p. 910). Here are some other examples. Masaccio who had "laid the groundwork for what Raphael perfected" was in Raphael's shadow from which Berenson finally removed him (Grove, 1996, vol. 20, p. 537). There has been "a quickening of interest in [Uccello's] work in the Table 4. Entry of Italian Artists into the Canon, 1568-1996 Grove Active Vasari Félibien Lanzi Burckhardt Berenson Chastel Grove Rank Name in Born Died 1550, 156 658-168 1789 1855 894-190 1956 1996 1 Michel Angelo (Buonarroti) F 1475 1564 2 2 2 2 2 2 2 2 Leonardo da Vinci F 1452 1519 2 2 2 2 2 2 2 3 Giotto di Bondone F 1267 1337 2 2 2 2 2 2 2 4 Raphael Sanzio C 1483 1520 2 2 2 2 2 2 2 5 Tiziano Vecellio V 1485 1576 2 2 2 2 2 2 2 7 Mantegna Andrea N 1460 1506 2 2 2 2 2 2 2 11 Botticelli Sandro F 1445 1510 2 1 1 2 2 2 2 12 Bramante Donato N 1444 1514 2 1 2 2 2 2 2 13 Giulio Romano C 1499 1546 2 2 2 2 1 2 2 14 Masaccio Tommaso F 1401 1428 2 2 1 1 2 2 2 15 Angelico Fra Giovani F 1395 1455 2 2 1 2 2 2 2 20 Verrocchio Andrea del F 1435 1488 2 2 1 2 2 2 2 21 Perugino (Vanucci) Pietro C 1450 1523 2 2 2 2 2 2 2 24 Ghirlandaio Domenico F 1448 1494 2 1 1 2 1 2 2 25 Lippi Filippino F 1457 1504 2 2 1 2 1 2 2 27 Salviati Francesco F 1510 1563 2 2 2 1 0 1 2 28 Sarto Andrea del F 1486 1530 2 2 2 2 2 2 2 31 Lippi Filippo F 1406 1469 2 2 1 2 2 2 2 32 Uccello Paolo F 1397 1475 2 1 1 1 2 2 2 33 Sebastiano del Piombo V 1485 1547 2 2 1 2 1 2 2 34 Bartolommeo Fra della Porta F 1472 1517 2 0 2 1 1 1 2 39 Carpaccio Vittore V 1460 1525 2 0 1 1 2 1 2 41 Bronzino Agnolo F 1503 1572 2 1 1 2 1 2 2 44 Parmigianino Francesco N 1503 1540 2 2 1 0 1 2 2 46 Cimabue Cenni di Pepo F 1485 1547 2 2 2 1 1 2 2 s 6 Tintoretto (Robusti) Jacopo V 1519 1594 1 2 2 2 2 2 2 8 Bellini Giovanni V 1431 1516 1 2 1 2 2 2 2 9 Veronese Paolo V 1528 1588 1 2 2 2 2 2 2 10 Giorgione (Ziorzi) V 1477 1510 1 2 2 2 2 2 2 16 Piero della Francesca C 1415 1492 1 2 2 2 2 2 2 18 Correggio (da) Allegri Antonio N 1489 1534 1 2 2 2 2 2 2 22 Martini Simone S 1284 1344 1 2 2 1 2 2 2 35 Antonello da Messina C 1430 1479 1 2 2 1 1 2 2 45 Castagno Andrea del F 1419 1457 1 2 1 1 2 2 2 29 Lotto Lorenzo V 1480 1556 1 1 2 1 2 2 2 40 Bassano Jacopo V 1510 1592 1 2 1 2 1 2 19 Francesco di Giorgio s 1439 1501 1 0 1 2 1 2 2 38 Signorelli Luca c 1450 1523 1 1 1 2 2 2 2 49 Pollai(u)olo Antonio F 1432 1498 1 0 1 2 2 2 2 17 Duccio di Buoninsegna S 1278 1319 1 1 1 1 2 2 2 26 Domenico Veneziano F 1405 1461 1 0 1 1 2 2 2 36 Tura Cosimo N 1430 1495 0 1 1 2 2 2 37 Gentile da Fabriano N 1385 1427 1 1 1 1 2 2 2 43 Roberti Ercole de N 1455 1496 1 1 1 2 2 2 48 Pisanello (Pisano) Antonio N 1395 1455 1 1 1 1 2 2 2 30 Masolino da Panicale F 1383 1435 1 1 1 1 1 2 2 42 Lorenzo Monaco F 1370 1425 1 1 1 1 2 2 50 Cavallini Pietro C 1 1 1 1 0 2 2 23 Bellini Jacopo V 1400 1470 1 1 1 1 1 1 2 47 Paolo Veneziano V 1333 1362 0 0 0 0 0 1 2 m TI O 3D ä 5 o z o TI o > z o z C/3 Note: All artists to whom Vasari devotes a vita or a note, who are discussed by Felibien and Lanzi, cited by Burckhardt, Berenson, Chastel, or in Grove's article on the Italian Renaissance are coded 2 (if among the top 50) or 1 (if present). Others are coded 0 (if absent). C: Center; F: Florence; N: North; S: Sienna; V: Venice. See also text, fl = floruit. Table 5. Entry of Flemish Artists into the Canon, 1604-1996 Grove Mander Sandrart Descamps Fiorillo Immerzeel Wurzbach Puyvelde Grove Rank Name Born Died 1604 1675 1753 1815 1842-55 1911 1953-70 1996 2 Breughel Pieter ? 1569 2 2 2 2 2 2 2 2 4 Weyden Rogiervan der 1400 1464 2 2 1 2 1 2 2 2 5 Eyck Jan Van 1381 1441 2 2 2 2 2 2 2 2 6 Bosch Hieronymus 1450 1516 2 2 1 2 2 2 2 2 7 Goes Hugo van der 1440 1482 2 2 1 2 2 2 2 2 8 Gossart (Mabuse) Jan 1478 1532 2 2 2 2 2 2 2 2 13 Metsys Quinten 1466 1530 2 1 2 2 2 2 2 2 14 Orlrey (Von brussel) Bernard 1488 1541 2 1 1 2 2 2 2 2 19 Mor van Dashorst Antonis 1516 1576 2 2 2 2 2 2 2 2 22 Cleve Joos Zotte ? 1541 2 2 1 1 1 2 2 2 23 Patinir Joachim 1480 1524 2 1 1 1 1 2 2 2 28 Vredeman de Vries Hans 1527 1606 2 0 2 1 2 1 0 2 31 Floris Frans 1519 ? 2 2 2 2 2 2 2 2 39 Veen Otto 1556 1629 2 2 2 2 2 2 1 2 42 Aertsen Pieter 1508 1575 2 0 2 2 2 2 2 2 45 Eyck Hubert Van 1385 1426 2 2 1 1 2 2 2 2 48 Coecke van Aelst Pieter 1502 1550 2 2 1 2 2 2 2 2 49 Mostaert Jan 1528 1598 2 2 2 2 1 2 2 2 50 Coninxloo Gillis 1544 1604 2 2 1 1 1 1 2 2 1 Rubens Pieter 1577 1640 0 2 2 2 2 2 2 2 3 Dyck Anthony van 1599 1641 0 2 2 2 2 2 2 2 10 Jordaens Jacob 1593 1678 0 2 2 2 2 2 2 2 16 Vos Marten 1532 1603 1 2 2 2 2 2 2 2 18 Janssen Abraham 1575 1632 0 2 2 2 1 1 1 2 21 Brouwer Adrian 1606 1638 0 2 2 2 2 2 2 2 47 Valckenborch Lucas 1535 1598 1 2 1 1 1 1 1 2 s 11 Memling Hans 1435 ? 0 1 2 2 2 2 2 2 20 Snyders Frans 1579 1657 0 0 2 2 2 2 2 2 24 Teniers David the younger 1610 1690 0 1 2 2 2 2 2 2 25 Thulden Theodoor 1606 1676 0 0 2 1 1 1 1 2 30 Breughel Jan Velvet I 1568 1625 0 1 2 2 2 2 2 2 38 Crayer Gaspar 1584 1669 0 1 2 2 2 2 2 2 46 Heem Davidsz. Jan 1606 1683 0 1 2 1 2 2 1 2 26 Hemessen Jan Sanders 1519 1556 0 0 0 2 0 1 2 2 44 Diepenbeeck Abraham 1596 1675 0 1 1 2 1 2 1 2 32 Quellinus Erasmus 1607 1678 0 1 1 1 2 2 1 2 33 Francken Frans II 1581 1642 0 0 1 1 2 2 1 2 12 Bouts Dieric I 1415 1475 0 0 0 0 0 2 2 2 15 David Gerard 1460 1523 0 0 0 0 0 2 2 2 17 Justus van Gent (Wassenhove) fl 14 20-40 0 0 0 1 1 2 2 2 29 Vellen (Velart) Dirk 1480 1547 0 0 0 0 0 2 1 2 35 Christus Petrus 1410 1476 0 0 0 1 0 2 2 2 43 Fyt Jan 1611 1661 0 0 1 1 1 2 1 2 9 Master of Flemalle 1375 1444 0 0 0 0 0 1 2 2 34 Breughel Pieter II 1565 1638 0 0 0 1 1 1 2 2 37 Isenbrandt Adrian ? 1551 0 0 0 0 1 1 2 2 40 Massys (Metsys) Jan 1509 1575 0 0 0 1 0 1 2 2 27 Vos Cornells de 1592 1667 0 0 0 0 1 1 1 2 36 Vos Paul de 1591 1678 0 0 1 1 1 1 1 2 41 Benson Ambrosius 1500 1550 0 0 0 0 0 1 1 2 m Tl O 3D o z o Tl o > z o z w Note: All artists discussed by van Mander, Sandrart, Descamps, Fiorillo, Immerzeel, Wurzbach, van Puyvelde and cited in Grove's article on Flemish painting are coded 2 (if among the top 50) or 1 (if present. Others are coded 0 (if absent), fl = floruit. 58 / GINSBURGH AND WEYERS 20th century, mainly because of its appeal to modern sensibilities" (Grove, 1996, vol. 31, p. 517). Carpaccio "stood in the shadow of the Bellini brothers, and it was not until John Ruskin's passionate appraisal of his work in the 1860s that [he] emerged as a fully autonomous artistic personality" (Grove, vol. 5, p. 822). Bronzino's reputation declined with the reaction against maniera. "Only since World War II has he been recognized as the most sophisticated and technically accomplished Italian painter to embody the ideals of maniera" (Grove, 1996, vol. 4, p. 859). Parmigianino's "reputation declined as his works became of less interest to practicing artists. There has been a new appreciation of his work in the 20th century" (Grove, 1996, vol. 24, p. 202). Cimabue temporarily lost some of his reputation after the Madonna di Rucellai, which had been attributed to him by Vasari, was reattributed to Duccio by Berenson. All of those who were singled out by Vasari can therefore also be considered to have qualified in terms of criterion (a), failing despite scrutiny, to reveal defects sufficient to be disqualified. This makes for 25 out of 50 artists who are present in the first circle of the late 20th century canon since 1568. Felibien The next group of nine artists entered at the time of Felibien. Recall that Vasari found North Italian and Venetian artists less perfect than Florentine painters. This is why Tintoretto, Veronese, Giorgione, Bellini, and Corregio were only introduced by Felibien, though Vasari devoted to each of them sometimes long and favorable (in the case of Tintoretto, Giorgione, and Corregio) vite or notes. Five among these nine artists (Tintoretto, Veronese, Giorgione, Piero della Francesca, and Corregio) never disappeared from the top group. The four others (Giovanni Bellini, Simone Martini, Antonello da Messina, and Castagno) left the canon at some moments (Castagno, in particular, was considered vulgar by Cavalcaselle and Berenson), but were never omitted from the various narratives. One may consider that all nine entered the canon following Silvers' criterion (b): previously unnoticed meritorious or agreeable properties sufficient to qualify them were discovered, or at least better underlined. Lanzi Lorenzo Lotto and Jacopo Bassano—both from Venice, the second is simply cited by Vasari, but discussed by van Mander and praised by Ridolfi, who "placed him among the great Venitians" (Grove, 1996, vol. 3, p. 351)—were introduced by Lanzi, but disappear again with Burckhardt. Both found the acclaim they deserved thanks to Berenson and Venturi (1967/1901), but none of them is a new discovery since Vasari and Felibien knew them. Silver's criterion (b) seems to apply again. ON THE FORMATION OF CANONS / 59 Burckhardt Similar arguments can be invoked for the three artists (Maurizio Francesco di Giorgio, Luca Signorelli and Pollaiolo) introduced by Burckhardt. Berenson Berenson introduces six additional artists at the dawn of the 20th century. Duccio appears obviously because of the famous Madonna di Rucellai that was attributed to him by Berenson. Domenico Veneziano has a vita (coupled with that of Castagno), but did not make it to the canon before Berenson. The Ferrarese artist Tura, who was forgotten 50 years after his death in 1495, is one of the rare artists who has neither a vita nor a note in the Vite, appears quite late with Lanzi, thanks to the Ferrarese art historian Baruffaldi (1675-1755) who restored his reputation, and considered him as the founder and greatest representative of the Ferrarese school. He became part of the canon with Berenson, after art historians Campari (1821-1887) and Venturi (1856-1941) rediscovered his lost works through archives (Grove, 1996, vol. 31, p. 433). Together with him appears Ercole di Roberti, also from Ferrara, and Pisanello who used to work in Ferrara, Verona, Mantua, and Milan. Gentile's case is somewhat different. He was "perceived as the consummate master of the late Gothic style [as] Masaccio alone was credited with the introduction of space and of light and shade into Renaissance painting" (Grove, 1996, vol. 12, p. 302). His role as a "progressive artist" was recognized only in the early 20th century. With the exception of Tura, all others are present in narratives since Vasari's time. Here we may be getting to the borderline between Silvers' criterion (b) of accessing the canon because of previously unnoticed meritorious properties, and criterion (c) of acquiring valuable properties sufficient to qualify. Chastel Though they were present in almost all narratives since Vasari, Masolino, Lorenzo Monaco, and Pietro Cavallini are canonized by Chastel, because new works by them were discovered or reattributed to them. This obviously qualifies them according to Silvers' criterion (b) of unnoticed, but existing, properties. Masolino's work had often not been distinguished from that of Masaccio, and it was only in the mid-19th century that his independent fresco cycle in Castiglione Olona was recovered, while 100 years later the frescoes and sinopie of another cycle were discovered in Empoli (Grove, 1996, vol. 20, p. 553). Likewise, in the mid-19th century, Crowe and Cavalcaselle reattributed to Lorenzo paintings that were thought to be by Giotto, Taddeo Gaddi, and other 14th century painters (Grove, 1996, vol. 19, p. 683). Cavallini, finally, was thought a pupil of Giotto, and this view prevailed until the early 20th century. It was reassessed after the discovery of the Last Judgment in Santa Cecilia in 1900, followed by more works attributed to him later (Grove, 1996, vol. 6, p. 107). 60 / GINSBURGH AND WEYERS Grove The two last cases are Jacopo Bellini and Paolo Veneziano. Bellini's Venetian works were lost very early, and neither Vasari nor his team could have seen them. An important addition to Bellini's work "came with the reappearance of the volume on parchment that was bought by the Louvre in 1884, which was followed by a wide range of critical writings. His high standing as an artist and his fundamental historical role were continually debated until the late 20th century when he received the recognition that is his due" (Grove, 1996, vol. 3, p. 654). Paolo Veneziano's very late appearance with Chastel (though Chastel does not rank him among the top 50) is due to the very recent realization that he was instrumental in merging Gothic and Byzantine art by quoting from both: "Understanding Paolo's art and that of Venice as a whole in this period has been hampered by a false dichotomy between Gothic and Byzantine influences and by the failure to appreciate the progressive role of Byzantine painting... [Paolo's] influence on later Venetian painters of the 14th century seems to have been fundamental and almost universal" (Grove, 1996, vol. 33, p. 33). Bellini is again clearly a Silvers' (b) case. Paolo's entry in the canon is more difficult to qualify, and could fall under Silvers' (c) criterion. It is remarkable that Vasari gave birth to half of the late 20th century first circle's canon, with artists satisfying Silvers' criterion (a). Narratives made them being, almost always, part of the canon. Félibien, who was keen to take on board North Italian artists, added nine names. Lanzi and Burckhardt added five; Berenson made room for six, of which Tura may have entered because of newly discovered properties, Silvers criterion (c). Chastel added another three to the list. For reasons described in some detail before, Silvers' criterion (b) is likely to apply to all of them, as well as to Jacopo Bellini. Paolo Veneziano should probably benefit from having entered the canon, thanks to Silvers criterion (c). There are thus two accessions to the canon for which criterion (c) could in principle have played a role—Tura and Paolo Veneziano—though it is hard to decide whether the reason is more ontological than epistemological. One may also appeal to criterion (c) for Uccello and Piero della Francesca who were praised for their "Cubist" manner. The French painter and art critic André Lhote (1930, cited by Del Buono, 2006, p. 169) thinks of Piero as the first cubist artist, and Clark (1983) "compares Uccello's achievement to that of George Seurat and likened his methods to those of the Cubists" (Grove, 1996, vol. 31, p. 517). Flemish Painters A similar picture emerges for the Flemish canon in Table 5, for which we follow the same presentation as the one for the Italian Renaissance, starting with van Mander (1604) and Sandrart (1675), following up with Descamps (1753, ON THE FORMATION OF CANONS / 61 1760-1764), Fiorillo (1815), Immerzeel (1842, 1855), Wurzbach (1906-1911), van Puyvelde (1953, 1962; van Puyvelde & van Puyvelde, 1970) and ending with the Grove Dictionary (1996). van Mander and Sandrart Six names (Pieter Breughel, Jan Van Eyck, Gossart, Metsys, Mor van Dashorst, and Floris) appear with van Mander, and are there to stay. Van der Weyden should also be part of this group had he not been attributed many unworthy pictures during the 18th and 19th centuries (Grove, 1996, vol. 33, p. 127). Similar comments apply to Joos van Cleve following the confusion between him and his son Cornells. Four additional painters (Rubens, van Dyck, Jordaens, and Brouwer) could hardly have been singled out by van Mander, since they were too young in his time, but they are added by Sandrart, and do not leave the first circle of the canon. So is Vos, who was already known by van Mander. Abraham Janssen leaves the canon during the 19th and 20th centuries, with a comeback in Grove's Dictionary. By the time of Sandrart, 26 names—that is, like in the Italian case, half of the contemporary canon's first circle—were already there in 1675. It may also be worth pointing out that out of the 23 artists of Lampson's (1956/1572) canon, who wrote 30 years before van Mander, and 100 years before Sandrart, 13 are present in the Flemish contemporary canon. Some disappear from time to time over the centuries: Sandrart fails to mention Aertsen. Neither Sandrart nor van Puyvelde mention Vredeman de Vries; this may be due to the fact that he was mainly involved in trompe l'oeil wall paintings and architecture and fortifications engravings. Little is known about his paintings. But narratives kept going for all the others and they are part of today's canon, most of them qualifying according to Silvers' criterion (a). Descamps, Fiorillo, and Immerzeel Sandrart already mentions five of the seven names (Memling, Snyders, Teniers the Younger, Thulden, Breughel Velvet, Crayer, and Heem) that enter next with Descamps. Thulden's work was dismissed in the beginning of the 19th century, but is "beginning to be accorded [its] rightful place in the history of art" (Grove, 1996, vol. 30, p. 789). Van Mander and Sandrart very briefly mention Hemessen, but "van Mander gave little information about him and failed to appreciate the avant-garde aspects of his work, characterizing him as an archaizing painter" (Grove, 1996, vol. 14, p. 381). He and Diepenbeek are rediscovered by Fiorillo in the early years of the 19th century only. So are Quellinus and Frans Francken II, who appear somewhat later with Immerzeel in the mid 19th century. 62 / GINSBURGHAND WEYERS Wurzbach A group of so-called Flemish primitives enters only in the beginning of the 20th century, some because more art historic research discovered them, some because their work had previously been misattributed. Friedländer's important essays on early Netherlandish painters (1903, 1924-1937, 1956/1916) are obviously no less important here than were those of Berenson.2 He introduces Petrus Christus, Gerard David, Dieric Bouts, Justus van Gent, and the Master of Flemalle. Petrus Christus, for example, suffered from the lack of evidence concerning his artistic origins. David's fame diminished after his death for unknown reasons. Dieric Bouts' work was attributed to Memling until 1833, Justus van Gent's to Pedro Berruguete (see Ainsworth, 1992). Though most of them were also known by Vasari, they were forgotten until the beginning of the 19th century, where they started to be collected by Melchior and Sulpiz Boisseree from Cologne, and later studied and published by James Weale, who played a significant role in the rediscovery of Flemish painters, organizing large exhibitions in Bruges in 1867 and 1902. This rehabilitation, explains Sulzberger (1961), "is the consequence of a more general interest in the Middle Ages which develops jointly in Germany, France, England and the Low Countries. The cause has enthusiast defenders, but also powerful detractors whose opposition is based on their prestige; the conflict is fuelled by a conflict of generations, since those who are in favor of the gothic are young people, anxious to oppose to well-established values" (p. 9). van Puyvelde Finally, seven names enter the first circle during the 1950s only. The most important is the Master of Flemalle, so christened by von Tschudi in 1898 from three surviving parts of a lost panel. He was obviously "unhelped by a personality cult" (Grove, 1996, vol. 20, p. 672). Today, he is identified with Robert Campin, to whom Wurzbach devotes some space, and does also very briefly mention Flemalle. Pieter Brueghel IPs works were merely "recognized as copies and imitations of his father's most famous compositions" (Grove, 1996, vol. 4, p. 910; 2 They were both considering themselves as privileging connoisseurship and enjoyment of art rather than art history. Here is what Friedländer (1956/1916, p. vi) writes about art historians: "The ability to attribute and check attributions will then follow automatically from study and enjoyment. Yes, from enjoyment! Many art historians, it is true, make it their ambition to exclude pleasure from art, in which, for obvious reasons, some of them succeed too well____Reasoning based on calculations and measurements is presented as the true method. A dry approach stands high in favor. Abstruseness, involved terminology, which makes the reading of art-historical books such torture, derives from that very ambition. Sometimes there are depths, but so obscure as to be worthless for the reader, generally all is shallow but cunningly troubled so as to suggest depths." ON THE FORMATION OF CANONS / 63 Friedländer, 1956/1916, p. 133), and his oeuvre was neglected in the 18th and 19th centuries. Finally, Isenbrandt's works were thought to be by David, and reattributed in 1902 only. Grove The three last introductions are due to the Grove (1996) Dictionary. Paul de Vos is mentioned by Descamps, Cornells de Vos, much later, and Benson only in the beginning of the 20th century, probably in the wake of the exhibitions organized by Weale in Bruges in 1867 and 1902. Except for those artists who have been the object of misattributions, it is hard to say whether their late rediscoveries (that is after the time of Sandrart) are due either to their merely technical rediscovery, for example by von Tschudi or Weale, or to revealing previously unnoticed properties, that is, Silvers' criterion (b) or to the acquisition of properties in the light of later artworks, that is, Silvers' criterion (c). CONCLUSIONS Half of the artists in the two contemporary canons that we study had already been introduced 400 years ago (Silvers' criterion (a)). Wrong attributions (Duccio or the Master of Flemalle) or new technical discoveries (Cavallini was thought for long to be a pupil of Giotto, while he preceded him) prevented some names to be canonical any earlier. Some artists entered the canon at later times, given that art historians learned to appreciate or to understand them better (Silvers' criterion (b) ). The group of artists who entered or were moved to the first circle because their works acquired new properties in the light of later works (Silvers' criterion (c) ) is small (Paolo Veneziano, Uccello, Piero della Francesca, and to some extent, Tura). One may argue that changes of attributions and new discoveries (the Flemish primitives) are contextual and should be included in the latter group. Even this would not make their number very large. Though criteria such as invention, originality, newness, and progress, and their relative weights in evaluating artists change over time, it is surprising that half of the canons were there almost from the beginning. This appears to be in contradiction with the suggestions made by Junod (1976), Genette (1994), and many others that canons are continuously moving and that no artist can survive forever. One reason for this apparent discrepancy may be that we examine two "closed" canons, that is, canons that make no room for artists who appeared after 1600, and for rather small (even if artistically important) regions of Europe. A contemporary canon that would be devoted to "European great painters of all times" would probably include Manet, Duchamp, Picasso, and Pollock, and certainly exclude Botticelli, not to speak about Pietro Cavallini, Thulden, Mostaert, and van Coninxloo. But Leonardo, Michelangelo, Van Eyck and 64 / GINSBURGH AND WEYERS Rubens would probably be there to stay forever. Our finite memory gives more importance to newness and forces out "those whose classical status is most tenuous" (Westphal, 1993). The closed character of the canons that we examine, as well as appealing to well-known historians, may explain why we witness less discoveries or rediscoveries and shifts than those that Haskell (1976) describes. His celebrated work is mostly based on the behavior of art collectors and much less on the opinion of art historians, and changes of taste to which well-trained art historians such as those on whom we based our research should be less prone. Botticelli never left the canon. He became indeed peripheral during 200 years after Vasari's description, but regained centrality with Burckhardt in his 1855 Cicerone. And so did many other Italian and Flemish painters who are the subject of this article. APPENDIX 1 Defining the Two Canons Italy Our interest centered on the parts of entry on Italian Art in Grove's Dictionary (1996) devoted to the following subsections of vol. 16, pp. 654-668: "Late medieval painting, c. 1100-c. 1400," "Early Renaissance painting, c. 1400-c. 1500," "High Renaissance and Mannerist painting, c. 1500-c. 1600." We collected all the names cited in these subsections. They appear in Table 1. Baldassare Peruzzi, Cima da Conegliano, and Federico Barocci have entries in the Dictionary whose length would qualify them to appear among the top fifty. They are not cited in the entry on Italian Art and therefore do not appear in the tables. Others, such as Pellegrino Tibaldi who are not cited either also have rather long entries, though they would not be ranked among the top fifty. Some names were excluded from our lists for the following reasons: 1. artists but not painters (Pietro Aretino, Leon Battista Alberti, Poggi Braccioloni, Brunelleschi, Leonardo Bruni, the della Robbias, Ludovico Dolce, Donatello, Marcilio Ficino, Francia, Lorenzo Ghiberti, Christoforo Landino, Nanni di Banco, Andrea Palladio, Angelo Poliziano, Marcantonio Raimondi, Jacopo Sansovino, Sebastiano Serlio, Vespasiano da Bisticci); 2. painters born less than 20 years before the publication of the Vite in 1568 who could hardly be described there (Michelangelo Caravaggio, Agostino, Annibale, and Ludovico Carracci, Lavigna Fontana, Galizia Fede, Marietta Tintoretto); and 3. others who were not artists, but supported them, such as the Medici, Niccolo Niccoli, etc.—Lazzarro Vasari who has a vita was excluded since he is not cited by any other historian; so was Giorgio Vasari, mainly because in his Vite, he devotes 31 pages to Raphael's life and 42 to his own. ON THE FORMATION OF CANONS / 65 Flanders The sections of interest in Grove's entry on Flemish Art are "Before 1400," "1400-c. 1550," and "c. 1550-1600." We ignored names of artists quoted in the subsections devoted to "manuscript illumination" (pp. 552-553 and p. 555), and "graphic arts" (pp. 555-556). See Table 2. Barthelemy Spranger, Karel van Mander, Gerard Horenbout, and Jan Kessel II have entries in the Dictionary whose length would qualify them to appear among the top fifty. They are not cited in the entry on Flemish Art and therefore do not appear in the tables. Some names were excluded for the following reasons: 1. in spite of being in the general entry, they have no personal entry in Grove's Dictionary (Lucas Achtschellinck, Thomas Willeboirts Bosschaert, Jan van der Asselt, Dieric Bouts II, Pierre de Bruxelles, Pieter van Coninxloo, Lodewijk De Dijster, Jehan de Gand, Anselm van Hulle, Godfried Maes, Theodoor Roeyermans); 2. they were artists but not painters (Francois Duquesnoy, Artus Quellinus); 3. they are cited in the general entry, but were mainly active in a foreign country (Jean de Beaumetz, Melchior Broederlam, Juan de Flandes, Master of Moulins, Master of the Parement de Narbonne, Michel Sittow, Justo Suttermans); 4. they are cited but are not Flemish artists (Maarten van Heemskerck); and 5. they were not artists (Justus Lipsius, etc.). We also attributed to Master of Flemalle the lengths of possible entries devoted to Robert Campin, since according to recent research, they are the same person, though there is also a small entry for Campin in Grove's Dictionary. Since he is cited in the entry on Flemish painting, he nevertheless appears in the lists. Note that we introduce in the Flemish canon some painters, such as Aertsen, Mostaert and Heem, who were born in the North, because they were attracted to Flanders, in the same spirit that made us discard Flemish painters who left for Italy or Spain. APPENDIX 2 The Canons' First Circles or Why 50? In this appendix we justify our choice of including 50 painters as first circle in both canons. We certainly do not want this first circle to become too large, since then, almost all the painters who are in the two Grove articles would be canonical. We decided to follow Cutting's (2006a) suggestion and try to find a cutoff value using Zipf s law, based on an empirical regularity observed in many fields (for the many domains of application, see http://www.nslij-genetics.org/wli/zipf), including the arts as shown by Cutting. The "law" shows that the frequency or 66 / GINSBURGH AND WEYERS absolute number of occurrences of a certain category of TV items (words in a language or in a given text, populations in cities, number of citations of artists, etc.), is related to the their ranks (1, 2, 3, etc.) in the category. The mathematical form that relates the two series of numbers can be written ylXf =C, (1) where yt is the frequency, or the number of occurrences of item i in a list, x, represents its rank, a and C are two parameters of the law that can be determined empirically by running the following linear in logarithms (log) regression: log yi = - a log xt + C. (2) Figures 1 (Italian painters) and 2 (Flemish painters) illustrate the relations between length of the entries in Grove's Dictionary (on thej-axis) and ranks (x-axis). The upper part of each figure illustrates equation (1), while the lower part illustrates equation (2). As can be checked, the upper parts of the two figures show that the relations are very non-linear. This is no longer the case in the lower parts of the two figures, at least up to a certain rank (represented on the horizontal axis). In both figures, it can be seen that the relation is roughly linear for ranks that are smaller than 1.70 (which happens to be the logarithm of 50), and drops afterwards. The choice of the cutoff point, that is, the number of artists who belong to the first circle of the canon, can be determined by searching for the approximate rank after which (2) ceases to be linear. The intuition is therefore to run regressions of model (2), by varying M, the number of observations (here, the number of artists), and looking for the highest fit, measured by the coefficient of correlation (R2). Here we chose M= 10, 20, 30, N, where N= 125 for Italian and 129 for Flemish painters. Table 6 shows the values of the correlation coefficients associated to different choices of M. As is readily seen, the choice of Mis 50 for Italian painters (R2 = 0.966) and should be M= 70 (R2 = 0.977) for Flemish painters if we followed strictly our choice criterion. To keep things symmetric, and since for Flemish painters, the correlation coefficient increases only slightly when going from 50 to 70 painters, we also chose M= 50. ACKNOWLEDGMENTS We are very grateful to two referees and to the editor of this journal for extremely useful comments. James Cutting's remarks as well as his impressive analysis of the Impressionist canon, and his use of the Zipf distribution, clarified our views on the necessity to justify the number of painters who form the first circle of the canon. We are also very grateful to Philippe Junod for careful comments on a previous version and to Didier Martens who helped us selecting art historians and artists to describe the Flemish canon. ON THE FORMATION OF CANONS / 67 Figure 1. Zipf charts, Italian painters. 68 / GINSBURGH AND WEYERS Figure 2. Zipf charts, Flemish painters. ON THE FORMATION OF CANONS / 69 Table 6. Determining the Number of Artists in the "First Circle" Coefficient of correlation of Eq. (2) Number of- artists Italian painters Flemish painters 10 0.944 0.885 20 0.948 0.932 30 0.958 0.957 40 0.962 0.967 50 0.966 0.971 60 0.954 0.975 70 0.943 0.977 80 0.940 0.974 90 0.936 0.960 100 0.911 0.916 125 0.819 — 129 — 0.861 REFERENCES Ainsworth, M. (1992). Implications of revised attributions painting. Metropolitan Museum Journal, 27, 59-76. Bazin, G. (1986). Histoire de l'histoire de I'art. De Vasari änos jours. Paris: Albin Michel. Berenson, B. (1926/1894, 1896, 1897, 1907). Les peintres italiens de la Renaissance (4 vol.). Paris: Editions de la Pleiade. This is the 1926 translation of Berenson's four volumes Venetian Painters of the Renaissance (1894), Florentine Painters of the Renaissance (1896), Central Italian Painters of the Renaissance (1897), North Italian Painters of the Renaissance (1907). Bourdieu, P. (1983). The field of cultural production, or the economic world reversed. Poetics, 12, 311-356. Bourdieu, P. (1996). The rules of art. Cambridge: Polity Press. Bozal, V. (1997). Pinturas negras de Goya. Madrid: Tf. Editores. Burckhardt, J. (1855). Le Cicerone: Guide de I'art antique et de I'art moderne en Italie (2 vol.). Paris: Firmin-Didot. Chastel, A. (1995/1956). L'art Italien. Paris: Flammarion. Clark, K. (1983). Paolo Uccello and abstract painting. In K. Clark (Ed.), The art of humanism. London: Murray. Coetzee, J. M. (2002). What is a classic? In J. M. Coetzee (Ed.), Stranger shores. Essays 1986-1999. London: Vintage. 70 / GINSBURGHAND WEYERS Cutting, J. E. (2006a). Impressionism and its canon. Lanham, MD: University Press of America. Cutting, J. E. (2006b). Mere exposure, reproduction, and the Impressionist canon. In A. Brzyski (Ed.), Partisan canons (pp. 79-94), Durham, NC: Duke University Press. Del Buono, O. (2006). Piero della Francesca. Paris: Flammarion. Descamps, J-B. (1753, 1760-1764). La vie despeintres flamands, allemands et hollandais. Paris: Chez Charles-Antoine Jombert, tome 1; Paris: Dessaint et Saillant, Pissot, Durand, tomes 2-4. Felibien des Avaux, A. (1967/1725). Entretiens sur les vies et sur les ouvrages des plus excellens peintres anciens et modernes; avec la vie des architectes. (Edited by Sir Anthony Blunt, Farnborough: Gregg Press). Fiorillo, J. (1815). Geschichte der zeichnende Künste in Deutschland und den vereinigten Niederlanden (4 vol.). Hannover: Bei den Gebrüdern Hahn. Friedländer, M. (1903). Meisterwerke der Niederländischen Malerei des 15. und 16. Jahrhunderts auf der Ausstellung zu Brügge. München: F. Bruckmann. Friedländer, M. (1924-1937). Die altniederländische maierei (14 vol.). Leiden: Sijthoff. Friedländer, M. (1956/1916). From Van Eyck to Breughel. London: Phaidon Press. (Originally published in German in 1916 as Von Jan van Eyck bis Brueghel). Genette, G. (1994). L'oeuvre de l'art (2 vol.). Paris: Seuil. Ginsburgh, V., & Weyers, S. (2006). Persistence and fashion in art: The Italian Renaissance from Vasari to Berenson and beyond. Poetics, 34, 24-44. Gombrich, E. H. (1969). In search of cultural history. Oxford: Oxford University Press. Grove (1996). The dictionary of art (34 vol.). J. Turner (Ed.). New York: Grove. Haskell, F. (1976). Rediscoveries in art. Some aspects of taste, fashion and collecting in England and France. London: Phaidon. Hume, D. (1965/1757). Of the standard of taste. In On the standard of taste and other essays. Indianapolis: The Bobbs-Merrill Company. Hutter, M., & Shusterman, R. (2006). Value and the valuation of art in economic and aesthetic theory. InV. Ginsburgh&D. Throsby (Eds.), The handbook of the economics of art and culture. Amsterdam: Elsevier. Immerzeel, J. (1842). De levens en werken des Hollandsche en Vlaamsche kunstschilders, beeldhouers, graveurs en bouwemeesters van het begin der vijftiende eeuw tot heden. Amsterdam: J. C. Van Kesteren. Immerzeel, J., Jr. (1855). De levens en werken des Hollandsche en Vlaamsche kunstschilders, beeldhouers, graveurs en bouwemeesters van het begin der vijftiende tot op de helft der negentiende eeuw. Amsterdam: Gebroeders Diederichs (Unchanged reprint, Amsterdam: B. M. Israel, 1974). Junod, P. (1976). Transparence et opacitě. Lausanne: L'Age d'Homme. Junod, P. (1995). Comment une oeuvre d'art devient un classique. In P. Gisel (Ed.), La selection (pp. 95-108). Lausanne: Editions Payot. Junod, P. (2002). Dans l'oeil du rétroviseur. Pour une histoire relativisté. Artibus et Historiae, 45, 205-221. Junquera, J. J. (2003). The black paintings of Goya. London: Scala Publishers Ltd. Lampson, D. (1956/1572). Les effigies des peintres célébres des Pays-Bas. Brussels: Desclée de Brouwer. ON THE FORMATION OF CANONS / 71 Lanzi, A. (1824/1789). Histoire de la peinture en Italie depuis la Renaissance des beaux-arts jusque vers la fin du 18e siecle (5 vol.). Paris: Seguin/Dufart. Levinson, J. (1990). Artworks and the future. In J. Levinson (Ed.), Music, art and metaphysics (pp. 179-224). Ithaca, NY: Cornell University Press. Levinson, J. (1996). Work and oeuvre. In J. Levinson (Ed.), The pleasures of aesthetics (pp. 242-273). Ithaca, NY: Cornell University Press. Milo, D. (1986). Le phenix culturel: De la resurrection dans l'histoire de l'art. Revue Francaise de Sociologie, 27, 481-503. Panofsky, E. (1971). Early Netherlandish painting: His origins and character. New York: Harper and Row. Rajagopalan, K. (1997). Aesthetic ideology: The case of canon formation. British Journal of Aesthetics, 37, 75-83. Sandrart, J. von (1675). Teutsche Academie der Bau-, bild- und Mahlerey-künste. Nürnberg (Neu gedruckt Nördlingen: Verlag Dr. Alfons Uhl, 1994). Silvers, A. (1991). The story of art is the test of time. The Journal of Aesthetics and Art Criticism, ¥9,211-224. Simonton, D. (1998). Fickle fashion versus immortal fame: Transhistorical assessments of creative products in the opera house. Journal of Personality and Social Psychology, 75, 198-210. Sulzberger, S. (1961). La rehabilitation des primitifs fiamands, 1902-1867. Bruxelles: Academie Royale de Belgique. Teyssedre, B. (1964). L'histoire de l'art vue du Grand-Siecle. Paris: Julliard. van Mander, K. (1604). Het schilder-boeck. Haerlem: Paschier Van Wesbusch. Van Peer, W. (1996). Canon formation: Ideology or aesthetic quality. British Journal of Aesthetics, 36, 97-108. van Puyvelde, L. (1953). La peinture fiamande au siecle des Van Eyck. Amsterdam: Elsevier. van Puyvelde, L. (1962). La peinture fiamande au siecle de Bosch et Breughel. Paris: Elsevier. van Puyvelde, L., & van Puyvelde, T. (1970). La peinture fiamande au siecle de Rubens. Bruxelles: Ed. Meddens. Van Rees, C. J. (1983). How a literary work becomes a masterpiece: On the threefold selection practiced by literary criticism. Poetics, 12, 397-417. Vasari, G. (1981/1568). Les vies des meilleurs peintres, sculpteurs et architectes. A. Chastel (Ed.). Paris: Berger Levrault. Venturi, A. (1967/1901). Storia dell'arte Italiana. Milano: Ulrico Hoepli (Reprinted by Kraus Reprints, Nendeln, Liechtenstein, 1967. Artists index compiled by Jacqueline D. Sisson, Nendeln, Liechtenstein, Kraus-Thompson Organization. 1975). Verdaasdonk, H. (1983). Social and economic factors in the attribution of literary works. Poetics, 12, 383-395. Verdaasdonk, H. (2003). Valuation as rational decision-making: A critique of Bourdieu's analysis of cultural value. Poetics, 31, 357-374. Westphal, M. (1993). The canon as flexible, normative fact. The Monist, 76, 436-449. 72 / GINSBURGHAND WEYERS Wurzbach, A. von (1906-1911). Niederländisches Künstler-lexikon (Reprinted Amsterdam: Boekhandel en Antiquariaat, B.M. Israel, 1963). Direct reprint requests to: Victor Ginsburgh ECARES CP. 114 Universitě Libre de Bruxelles 50 avenue F. Roosevelt 1050 Brussels, Belgium e-mail: vginsbur@ulb.ac.be