The tradition of reasoning about legitimate authority, law and community from the Quran, the practice of the Prophet and the accumulated body of Islamic jurisprudence. Its organising question is the relationship between revealed law and human government: whether rulers derive authority from applying sharia, from consent, from lineage, or from the simple necessity of order.
The classical tradition is more varied than modern polemic suggests. Al-Farabi transposed Plato's philosopher-king into an Islamic virtuous city; al-Mawardi wrote a legalist account of the imamate that conceded much to the political facts of his day; al-Ghazali argued that even flawed rule is preferable to civil strife; Ibn Taymiyya pushed toward a stricter subordination of ruler to law; and Ibn Khaldun, standing somewhat outside the genre, explained the rise and decay of dynasties through asabiyyah, group solidarity, in what reads as sociology centuries before the discipline existed.
Modern Islamism is a response to colonial subordination rather than a simple continuation. Al-Afghani and Abduh sought to reconcile Islam with modern science and constitutional government; Rashid Rida moved toward restoring the caliphate; Mawdudi introduced hakimiyya, God's exclusive sovereignty, which Sayyid Qutb radicalised in Milestones (1964) by declaring existing Muslim societies to be in a state of pre-Islamic ignorance and therefore legitimate targets. Khomeini's velayat-e faqih — rule by the jurist — was a Shi'i innovation without real precedent, contested by senior clerics who held that clerical rule corrupts both religion and government.
The Islamist current is a minority position within a much larger field. Constitutionalist and democratic strands run from the Ottoman and Iranian constitutional movements to Ghannouchi's argument that democracy is the appropriate modern form of consultative rule, and the quietist majority of jurists have historically kept clerical authority out of government entirely. Reading the tradition as though Qutb were its natural endpoint misdescribes fourteen centuries of argument.