DEVELOPMENT OF THE HISTORY OF ART 583 



finality in itself. It needs support from without, and every scrap of 

 evidence that corroborates should be brought to bear. 



As for evidence itself and its weight I sometimes sigh for a good 

 book on the " Value of Human Testimony," and a companion volume 

 on "What is Logic?" They should be in 1 the hands of every historian 

 of art. It is necessary, of course, that the connoisseur should know 

 what is a copy, what a variant, what an original; but it is also 

 necessary that he should know what is common sense. It is not, for 

 instance, common sense to cast out all documents about pictures 

 or marbles simply because some of them have been misleading cr 

 erroneous. A Raphael contract or agreement to paint a Hercules and 

 the Nemean Lion may be worthless because the agreement was never 

 carried out; but a Raphael agreement for a "School of Athens" 

 would be excellent evidence because the agreement was carried out. 

 To be sure, a document may point to a certain altar-piece which 

 was afterward stolen and a copy quietly put in its place, and in such 

 a case criticism is justified in saying that the copy is a copy and not 

 the original; but the agreement of Correggio to paint the "Holy 

 Night " now in the Dresden Gallery is extant and is good corrobor- 

 ative proof of the Dresden picture having been painted by Correggio. 

 True enough documents have been forged and so also have signatures 

 forged galore but there are true documents as there are true 

 signatures, and either or both may be trustworthy evidence. The 

 question of probability comes in just here. There is nothing inher- 

 ently improbable about the inscription on the St. Bavon altar-piece 

 to the effect that Hubert van Eyck began it and Jan van Eyck 

 finished it. If it were a lie, it would not have been tolerated there 

 in the first place. It has always been accepted as a true statement 

 until the recent exhibition of early Flemish art at Bruges gave the 

 critics a chance to spin theories and formulate doubts. The St. 

 Bavon altar-piece failed to fit the theories and, of course, the theories 

 could not be in error. The altar-piece was wrong. Then followed slur 

 and innuendo, the glance askance, and the "I could an I would," 

 all because the critics wanted to reconstruct the lost personality of 

 Hubert van Fyck by taking away from the established personality 

 of Jan van Kyck. In fact the defects of the newer criticism have been 

 exemplified in the most extravagant form in the recent attempts at 

 rewriting the history of the early Flemings. The writers have put 

 down a long series of unsupported guesses and asked their acceptance 

 as facts, ignoring all papers, past histories and traditions as mere 

 "petty documentation." 



Without doubt a signature or inscription needs support by the 

 internal evidence of the work itself, but where one confirms the other 

 both should be accepted. And every one knows that written history. 

 such as that of Lucian or Vasari. is not to be trusted implicitly. It 



