Part of the reason Bayesian analysis was avoided was because, prior to Pearl's work, inference was impractical in non-toy examples. A major contribution of Pearl's work was to make it feasible, by structuring the probability distributions as Bayesian networks that limited the possible dependencies, coupled with the belief-propagation algorithm to do approximate updates.
I don't see Pearl as primarily interested in the Bayesian v. frequentist debate himself, though, but rather in how to efficiently do probabilistic reasoning in non-trivial problems in general, with a heavy tilt towards questions of representing causality. Methodologically his work over the years has used all sorts of things from various camps; for example, he was also an authority in the early 1980s on heuristic search.
I don't see Pearl as primarily interested in the Bayesian v. frequentist debate himself, though, but rather in how to efficiently do probabilistic reasoning in non-trivial problems in general, with a heavy tilt towards questions of representing causality. Methodologically his work over the years has used all sorts of things from various camps; for example, he was also an authority in the early 1980s on heuristic search.