The central tenet of science is that theories should be testable, and will only replace existing theories when their predictions are demonstrably better, or when they provide the same predictive power with fewer initial assumptions.
New theories (and their proponents) compete: as with evolution, the process of science selects for the fittest. Where selection pressure is low, as when two or more theories make rather similar predictions, competitors may coexist for a while. Sometimes they ultimately turn out to be logically equivalent: more often, new data emerges that contradicts one or other, which is then abandoned.
The evidence that this process works is everywhere, from miracles of modern communication to our ability to manipulate the genome. It only fails when science itself is abandoned or suppressed. The suppression of evolutionary theory, and promotion of creationism, in the US is a prime example, and would be fostered by the proposals for the revised National Curriculum which Wolpert decries.
The history and sociology of science - which theories arise when, and why - are rightly the province of historians and sociologists. It is a pity that some of their wilder ideas about the practice of science - how we distinguish between theories and invent new ones - are taken more seriously than those of practitioners.