Abstract: “Legal origins” scholars explain economic performance by a country’s membership in a given “legal family.” To demonstrate the proposition, they regress various indices of performance on, inter alia, that membership. These regressions are properly specified only if (a) countries cannot switch families, and (b) family membership seriously constrains legal change. If countries can switch, then family membership is endogenous to economic performance—since a country will decide whether to stay in a family with an eye to its expected economic effect. If countries can readily borrow across legal family lines, then membership does not bind—and necessarily can have no effect on performance. Unfortunately, neither of these propositions is true. Countries can indeed switch and borrow—easily. That one does not observe much cross-family switching or borrowing in practice merely reflects the fact— nicely demonstrated by Spamann—that countries find it easier to borrow from other countries that use the same language, and that legal families tend to correlate with linguistic families. Given that statutory options within any one legal family usually offer countries all the options they need, countries have little reason to move outside those linguistic groups. I illustrate the possibility of cross-family switching and borrowing with the example of pre-war Japan.