The publication of the first volumes by the committee of the Rembrandt Research Project (RRP) of Amsterdam has drawn renewed attention to the attributional debate surrounding the paintings. Like the drawings, the paintings can in general only be judged by visual, stylistic criteria, backed up by the study of X-radiographs, pigment analysis and other scientific processes. To take an extreme example, the Anatomy Lesson of Dr Tulp is believed to be by Rembrandt both because it ‘looks’ right and because there is nothing dubious about its physical components. If it failed to match our expectations of Rembrandt’s handling, quality and materials, notions of which are provided by other paintings attributed to Rembrandt and from the same period (for the most part without independent documentation), it would not be considered authentic. Signed and dated 1632 (although the inscription’s trustworthiness has been questioned by the RRP), the painting is first mentioned in a description of the city of Amsterdam published in 1693, more than 60 years after it was painted and 24 years after the artist’s death – hardly a confidence-inspiring source.
The RRP rightly accepts the painting among the genuine works it has catalogued so far, and although its findings and methods have come under attack, the central question as to whether it has dismissed too much from Rembrandt’s oeuvre revolves around only about 10% of the paintings the committee discussed (with a few more extreme dissenters). As compared to Horst Gerson’s 1968 catalogue that has been reduced by 30% or more, the RRP’s achievement looks positive. The Dutch scholars involved have also clarified the oeuvres of Rembrandt’s pupils, many of whom, including Gerard Dou, Govert Flinck and Ferdinand Bol, built solid reputations of their own in the 17th century.
Parallel investigations into the master’s drawings in recent years have yet to lead to a new catalogue raisonné. Perhaps for this reason they have attracted less publicity. Yet many specialists would agree that Benesch’s catalogue should be reduced in size by more than 25%. While the Benesch corpus remains fundamental to current research, it has been supplemented and superseded by the early reviews mentioned above and, in more recent years, by the catalogues produced by the major permanent repositories of Rembrandt’s drawings in Amsterdam (1985), Rotterdam (1988), Paris (exhibitions in 1988–9 and in 2006–7), London (Exh. 1992), New York (Exh. 1995–6), Paris (Institut Néerlandais, Exh. 1997–8), Hamburg and Bremen (Exh. Bremen, 2000–01), Munich (Exh. 2001–2), Vienna (Exh. 2004), Brussels (Exh. 2005), Braunschweig (Exh. 2006) and Berlin (2006). While these have relied heavily on Benesch’s work they make substantial inroads into his conception of Rembrandt’s draughtsmanship, rejecting between 20% and, in the case of Rotterdam, more than 50% of the works that Benesch had authenticated. In addition, the work of some of Rembrandt’s pupils and other related figures has been the subject of monographs and exhibition catalogues.
As mentioned above, the basis for current research must involve testing the plausibility of each attribution against a core of unquestioned, ‘documentary’ drawings. This method has been adopted in the past but not adhered to with rigour: the reliability of the core group has often been dubious. It should consist only of those drawings that have enjoyed an almost uninterrupted consensus eruditorum in their favour and which can be given to Rembrandt by virtue of connections that go beyond the confirmation provided by stylistic analysis; they must be related to other, if possible, documented works. Because scholars have improved opportunities for travel between the main repositories of Rembrandt’s drawings and more complete photographic archives at their disposal, they now have a better chance of success in isolating those drawings that can be reliably considered authentic. These drawings include the indented studies for Rembrandt’s etchings that have already been mentioned ( Cat. nos 5 and 31), or those with some other close connection to well-documented or generally accepted works. For example, the Study for the Portrait of Maria Trip (Cat. no. 23), as discussed in section 2, can be related to changes made by Rembrandt as he worked on a signed and dated painting, as revealed by X-radiography. Thus its attribution to Rembrandt is reasonably secure, as well as being confirmed on stylistic grounds. Reliably signed drawings can also usually be trusted, as discussed above.
Around a core of works of this kind, a few hundred drawings can be attributed to Rembrandt with some confidence on grounds of style and technique. Although not exhaustive, this expanded group gives a broad overview of Rembrandt’s draughtsmanship at most periods of his career. Sketches on the versos of several of these drawings expand the repertory of styles that can be accepted; but if a drawing fails to match any of those in this sizeable body of material, or if it appears to belong to a separate group, then questions have to be raised about its authorship. Such questions have often led to attributions to other artists from the master’s immediate circle, such as Raven and Drost. Beyond the core ‘documentary’ group the materials (paper, ink, chalk, etc.) and style are the only guides to further attributions. Simplistic attempts to weed out drawings of indifferent quality or to create a false, wholly infallible Rembrandt should be deprecated, and many of the drawings accepted today are not especially alluring or impressive when isolated from the context of his work as a whole. A few that were rejected on these grounds by Benesch and other writers have recently been reinstated.
The present reassessment of the British Museum’s Rembrandt drawings has led to a reduction from the 106 accepted by Benesch to 70 (excluding 4 that Benesch himself rejected, forgot or did not know), which is about 35% of the collection. Some are retained only with reservations. The collection, nevertheless, remains as large as or larger than any other in existence.
Over a longer historical perspective, the percentage of ‘losses’ is much greater. In 1858, Thoré-Bürger believed that Rembrandt drew some 150 sheets in the British Museum at that time, and because more than 40 that he would have probably accepted have since entered the collection, he would have accepted around 190. In 1899, four years after the arrival of the Malcolm collection, 98 of the Museum’s drawings were exhibited and catalogued by Sidney Colvin as by Rembrandt. Hofstede de Groot’s catalogue raisonné of 1906 accepted five more, excluding another 15 in George Salting’s bequest of 1910. This figure was more or less accepted, with just a few depletions by A.M. Hind, whose catalogue of the collection appeared in 1915. Benesch’s reduction to 106 drawings in the 1950s and today’s to 70 therefore belong to a slow and laborious process.
History strongly suggests that further changes will take place in the future, probably reducing the number of accepted drawings still further. In these circumstances a dogmatic or authoritarian approach, stating opinions without supporting arguments, is unwise. As long as opinions have to be formed on the basis of just a few well-documented drawings, there will always be areas of disagreement. Some progress should be within the reach of every generation of scholars, but there may yet be some truth in the opinion of the abbé Dubos who early in the 18th century, with a typically Gallic disrespect for the quacks of his time, wrote that ‘L’expérience nous enseigne que l’art de deviner l’auteur d’un tableau en reconnaissant la main du maître, est le plus fautif de tous les arts, après la médicine’.