The year 1958 was a busy time for Duchamp in England. British artist, Richard Hamilton, had proposed to Duchamp that he create a typographical version and translation of the famous Green Box Notes. Duchamp had visited Roland Penrose's house and knew him very well (R. Hamilton, personal communication, Fall, 1997).⁴ In the meantime, Hamilton himself was often at Roland's home. Lionel and Roger Penrose enter the story at this point. Tony Penrose (personal communication, Fall, 1997), Roland's son, testifies that Duchamp was at their home on more than one occasion.⁵ More significantly, Roger, and especially Lionel Penrose, were often at Roland's as well, playing chess and engaging in lively intellectual conversations. According to Tony Penrose, discussions of optical illusions, a subject that greatly interested both Roland and Lionel, inspired them to treat the topic in their writing (Penrose, 1973).

Thus, as the detective announces before confronting the suspect in a murder mystery, we have motive, means and opportunity. Given Lionel and Roger Penrose's shared interest (also held by Roland) in visual illusions (motive); their frequent meetings (opportunity); and Roland's apparent enthusiasm for Duchamp's work (means); it is likely that Roland Penrose showed Lionel and Roger the Apolinère Enameled work before or at the time of the 1958 publication of their discovery. If Lionel and Roger had, in fact, seen the bed in Apolinère Enameled before their publication, two interpretations seem plausible: (1) they saw the bed in 1958 but did not notice its status as an overt example of an impossible figure before they wrote their article; or (2) they noted the bed, talked about the phenomenon of ambiguity in dimensional representation and then devised and published the general category of impossible figures. I vote for the second. In a recent conversation, Roger Penrose told me that he was familiar with the idea that Apolinère Enameled was an impossible figure, but did not remember when he first recognized this (personal communication, Fall, 1997). Tony Penrose agrees with me that the second scenario seems more likely, and that his father probably discussed Duchamp's optical illusions with Lionel and Roger in the course of the brandy and chess conversations that often took place in his family home.

I have no smoking gun, but all the circumstantial evidence leads to the conclusion that Lionel and Roger Penrose's scientific discovery may have been influenced by the artist Duchamp. The dates speak for themselves. Duchamp deliberately used a distinctive example of an impossible figure in 1916-1917. Duchamp's knowledge and artistic expression of the phenomenon of ambiguity in dimensional representations thus occurred some forty years before the Penrose article. Moreover, the year that the Penroses published their paper was the very same year that they probably had seen Duchamp's bed at Roland Penrose's house. This fact, however, does not detract from the importance of the Penroses' discovery. Duchamp's demonstration provided one example of an impossible figure.⁶ The Penroses joined this with an entire category of optical illusions and coined the term "impossible figure." My point here is only to suggest one possible influence of an artist's work upon a scientific discovery. The more typical course of influence runs the other way, from science to art, as is well documented in art history.