Refutation of the claim that analogy operates only in matters of the Sacred Law and not in rational inquiries:

The later dialectical theologians and rationalists—such as Abū al-Maʿālī, Abū Ḥāmid, al-Rāzī, Abū Muḥammad al-Maqdisī, and others—who assert that analogy (qiyās) has no place in rational sciences and that it is confined to questions of the Sacred Law, maintaining that rational matters rest solely on whatever evidence points to the conclusion in an unrestricted way, stand in opposition to the position of the majority of Muslim theoreticians and, indeed, of all rational people.

Analogy is invoked in rational disciplines just as it is in the sciences of the Sacred Law. Whenever it is established that a shared attribute necessitates a given ruling, that serves as proof across all fields of knowledge; likewise, when it is established that no relevant distinction exists between the branch and its original case, that too is proof in every science. Wherever analogical reasoning by representation (qiyās tamthīlī) is inadmissible, analogical reasoning by subsumption (qiyās shumūlī) is likewise inadmissible.

Abū al-Maʿālī and the speculative theologians before him neither follow nor approve the method of the logicians; rather, they put forward arguments which they believe entail their conclusions without attending to formal structure. The logicians, on the other hand, and the majority of theoreticians, reason from the known to the unknown whenever the shared element necessitates the ruling, illustrating this by uniting things under a single definition, cause, condition, or evidentiary principle.

Their opponent objects, saying that the ruling concerning the unseen is not established because it holds in the observed; rather, the universal proposition itself suffices, and there is no need for such comparison. We reply: the same holds true in matters of the Sacred Law. Whenever evidence shows that the ruling is attached to the unifying attribute, there is no need to cite the original case; the very proof that the ruling hinges on that attribute is enough. Yet because a universal exists only insofar as it is instantiated in particulars, identifying an original case makes it possible to verify that universal. This is beneficial in both rational and revealed disciplines. Thus, wherever evidence demonstrates that the common attribute is the operative cause of the ruling, or that the difference between the original and the branch is legally irrelevant, the analogy is sound and constitutes valid proof, whatever the field.

Disagreement over what the term “analogy” denotes:

People have differed over the exact meaning of the word “analogy.”