Before he took over the management of the Economist in 1860, Walter Bagehot had not had much occasion to notice the United States, at any rate in his published writings. During the 1850s he had been too taken up with banking, and literary criticism, and expounding the value of stupidity in politics. To be sure, in 1859 he decided that the time had come to discredit the American example. The English were becoming disquietingly interested in democracy, a system as to which he had all the usual mid-Victorian doubts and a few extra. So he told the world, through the National Review (which struggling Unitarian quarterly he edited) that the vulgar American voters sent only vulgar men to Congress: “ men of refinement shrink from the House of Representatives as from a parish vestry ”; and that America was too unlike England to be a safe model. Then, just as he became editor of the Economist, the secession crisis and the Civil War erupted. It was incumbent on him to pronounce on these events, and it would have been most uncharacteristic of this sunny, self-confident man to shirk such a responsibility.
Uncertainty-aware Contact-safe Model-based Reinforcement Learning
