Shea snapped his fingers for attention. «Just a second, Doctor. In a world such as you’re conceiving, would the laws of magic work because people believed in them, or would people believe in them because they worked?»

Chalmers put on the smile that always accompanied his intellectual rabbit-punches. «That question, Harold, is, in Russell’s immortal phrase, a meaningless noise.»

«No, you don’t,» said Shea. «That’s the favourite dodge of modem epistemologists: every time you ask them a question they can’t answer, they smile and say you’re making a meaningless noise. I still think it’s a sensible question, and as such deserves a sensible answer.»

«Oh, but it is meaningless,» said Chalmers. «As I can very easily demonstrate, it arises from your attempt to build your — uh — conceptualistic structure on an absolutistic rather than a relativistic basis. But I’ll come back to that later. Allow me to continue my exposition.»

«As you know, you can build us a self-consistent logic on almost any set of assumptions —»

Bayard opened his half-closed eyes and injected another sharp observation: «Isn’t there a flaw in the structure there, Doctor? Seems to me your hypothesis makes transference to the future possible. We should then become aware of natural laws not yet discovered and inventions not yet made. But the future naturally won’t be ignorant of our method of transference. Therefore we could return to the present with a whole list of new inventions. These inventions, launched into the present, would anticipate the future, and, by anticipating, change it.»