In the pre-capitalist period, religions that decisively determined culture and world-view designated the limits of the various phases of the "social metabolism" in a value-preferential, generally negative, less frequently positive manner.1 Areas which later became so important in the religiously prescribed regulation of approved social conduct, such as economics and political science, were seldom objectified into rationally conceptualised disciplines, and in fact we find in Islam neither a coherent economic theory nor an independent political discourse. Society was regulated by sharíca, the religious law of divine origin, which could not be essentially changed but only interpreted, and so the state had no independent existence as its most important sphere of responsibility, the making of laws, not only could not develop but was specifically forbidden. Since state and government – at least in theory – are derived from sharíca, Jesus' reply to the Pharisees "Render unto Caesar the things that are Caesar's; and unto God the things that are God's"2 is inconceivable in Islam, where, at least according to the ideal ruling, Caesar (caliph, sultan etc.) means what the religious law considers correct. That is to say, there is no difference between religion and state/polity or between society and state. This is why no independent political science has been able to take shape; it is actually embedded in thought about Islam and the regulation of the religiously correct organisation of co-existence with Islam in society, and the constantly growing number of problems that this presents. 3 This too works on the basis of the principle of maqām, "theme and variations", that is, the main theme is concealed in an excess of variation. A political theory of Islam sui generis is therefore not formulated as such (just as we find no theories of economics4 or education5). Naturally, we find plenty of political practices in the turns of events of Muslim imperial formation (this social variation provides the social historical material for the commonest depiction, the so-called "history of the Arabs" genre – Marx, in his time, called it "the storms of the political cloud-region"), but we must extract the Islam-determination of this practice from other sources. Basically, we find systematic or occasional descriptions in scholars in three fields; these, however, seldom address Sein and the constant changes in its always imperfect practice, which gives little cause for satisfaction (perhaps only Ibn Khaldūn does that in his description of the power-state), but they speak rather to Sollen, and they try again and again to conjure up the idealised (by the Prophet and the "caliphs that walk the true way") condition of the beginnings that is regarded as the standard, the norm, and the way to be followed. It is customary to distinguish three types of sources that can contribute to the absent political theory, and in fact some of the works on the political thought of classical Islam analyse these.6 These three sources are: the compendiums of the experts on religious law (fuqahāc) – based on the Koran, the prophetic tradition and the practices of early Islam, in which, broadly speaking, the theory of the exercise of dominance under religious law is expounded, reacting in the longish term to changes in historic circumstances; the colourful supply of "king-mirrors" to be found in adab literature, which rely heavily on Middle Persian traditions and set before the ruler in moralising tone the possible and permissible practice of the power of a king in possession of a divine mandate; and finally the political theories of the philosophers, which further develop the Hellenistic inheritance and continue in Islamic dress the Platonic legacy of what is expected of the "philosopher-king" (principally, knowledge and justice) and the "virtuous (ideal) city" (al-madīna al-fādhila) as a kind of utopia.7 It is clear that we can best learn details of the political power in question indirectly, and compile a coherent picture of one field at a time/alone (e.g. taxation8). Let us examine the individual types of sources and what they may have to offer a little more closely.