Abstract

Many spaces that need to be searched in the sciences are too unwieldy for random search to stand any hope of success. Success instead requires a nonrandom search. But how does one find a nonrandom search that stands a good chance of success? Even to pose the question this way suggests that such nonrandom searches do not magically materialize but need themselves to be discovered by a process of search. The question then naturally arises whether such a “search for a search” is any easier than the original search. This paper establishes a conservation of information theorem according to which the information required to find a successful search is always at least as large as the information required to successfully complete the original search. This result shows that information, like money, obeys strict accounting principles, leaves a trail, and can only originate from a prior information source.