In the United States, First Amendment protections for free speech are deeply associated with democracy. The dominant view in the Supreme Court and among commentators is that the primary (albeit not necessarily the only) reason we protect free speech is because of its essential role in advancing democratic self-governance. What are the implications of the democratic self-governance theory for free-speech protections outside the United States, in particular in nondemocratic countries? If we assume that the role of free speech is to advance democratic politics, then presumably non-democratic countries would have no reason to protect or tolerate speech. After all, if one rejects western-style liberal democracy, presumably one also rejects the subsidiary rules that undergird that system of government. The truth, however, is more complex. First of all, the vast majority of constitutions in the world grant at least some level of written protection for free speech, even though many of these constitutions are in countries which do not even purport to be free, multiparty democracies. Of course, many of these constitutional protections are shams; but it is simply not the case that no autocratic regimes permit free speech. The purpose of this paper is to explore how and why that might be so, and to consider whether the answers to these questions might have implications for domestic law.I begin by surveying the scope of global protections for free speech in written constitutions, and then examining in some detail three case studies of autocratic countries which have provided a degree of room for free speech: modern Communist China, the Soviet Union during the Glasnost era under Mikhail Gorbachev, and modern Qatar. In each case, I demonstrate that the regime provides meaningful protections for free speech, albeit with clear limits. I also argue that in each of these cases, the leadership has absolutely no interest in advancing democracy or surrendering their monopoly on power. Yet even without democracy, they perceive that permitting some degree of free speech advances their interests and the interests of their citizens and nations.Based on my case studies, I identify three distinct reasons why autocratic leaders might have an interest in permitting some freedom of speech by citizens. The first, and most significant, is internal control. In any large, bureaucratic system, central leadership often faces great difficulty in getting local officials to advance central policies and follow central leadership. Citizens can play an important role in identifying, and publicizing, corruption and lawlessness, as well as violations of central policy, at the local level. Second, free speech can act as a safety valve. Permitting some degree of free speech can, therefore, alleviate pressures for political change. Third, free speech as a form of citizen participation in government can lend legitimacy to a government, even without the legitimacy conferred by popular consent through elections. I also explore the countervailing factors – notably the desire for rulers to maintain their power – which result in clear limits on what sorts of speech will be tolerated in autocracies.I close by considering whether these alternative justifications for protecting free speech have any implications for speech within the United States. I argue they do, for this reason: even though our system of government is at its base democratic, actual citizen interactions with the government often are not experienced this way. This means that in addition to protecting democratic government, free speech also plays some of the same roles in the U.S. as in autocracies: permitting oversight over the bureaucracy, providing a safety valve, and granting legitimacy to high officials. I close by considering some doctrinal implications of this insight.