The most gigantic of these experiments was the French Revolution. To find out that a society is not to be refashioned from top to bottom in accordance with the dictates of pure reason, it was necessary that several millions of men should be massacred and that Europe should be profoundly disturbed for a period of twenty years. To prove to us experimentally that dictators cost the nations who acclaim them dear, two ruinous experiences have been required in fifty years, and in spite of their clearness they do not seem to have been sufficiently convincing. The first, nevertheless, cost three millions of men and an invasion, the second involved a loss of territory, and carried in its wake the necessity for permanent armies. A third was almost attempted not long since, and will assuredly be attempted one day. To bring an entire nation to admit that the huge German army was not, as was currently alleged thirty years ago, a sort of harmless national guard,[15] the terrible war which cost us so dear had to take place. To bring about the recognition that Protection ruins the nations who adopt it, at least twenty years of disastrous experience will be needful. These examples might be indefinitely multiplied. [15] The opinion of the crowd was formed in this case by those rough-and-ready associations of dissimilar things, the mechanism of which I have previously explained. The French national guard of that period, being composed of peaceable shopkeepers, utterly lacking in discipline and quite incapable of being taken seriously, whatever bore a similar name, evoked the same conception and was considered in consequence as harmless. The error of the crowd was shared at the time by its leaders, as happens so often in connection with opinions dealing with generalisations. In a speech made in the Chamber on the 31st of December, 1867, and quoted in a book by M. E. Ollivier that has appeared recently, a statesman who often followed the opinion of the crowd but was never in advance of it--I allude to M. Thiers--declared that Prussia only possessed a national guard analogous to that of France, and in consequence without importance, in addition to a regular army about equal to the French regular army; assertions about as accurate as the predictions of the same statesman as to the insignificant future reserved for railways. 4. REASON In enumerating the factors capable of making an impression on the minds of crowds all mention of reason might be dispensed with, were it not necessary to point out the negative value of its influence.