It was polished for them by the Muslim dialecticians.

The equitable critic’s reply to their objections against analogical inference:

They draw a distinction between the syllogism of inclusion (qiyās al-shumūl) and representative analogy (qiyās al-tamthīl), claiming that the first can yield certainty whereas the second yields only conjecture. This is a false distinction: wherever one of the two provides certainty, the other provides certainty as well, and wherever one affords no more than probability, the other likewise affords no more than probability. Whether a proof imparts certainty or probability depends not on the outward form in which it is cast, but on whether it contains something that necessarily entails the ruling. If either of the two includes a factor that infallibly necessitates the ruling, certainty is obtained; if it contains only what renders the ruling probable, the result is mere probability.

What in one pattern is called the “middle term” is, in the other, the “shared attribute.” The major premise, which affirms the adherence of the major term to the middle, is simply a statement of the causal efficacy of the attribute common to both the original case and the branch. Whatever establishes the truth of that major premise simultaneously establishes that the shared attribute necessitates the ruling; thus the attachment of the major term to the middle is the attachment of the ruling to the shared attribute.

For example, when you say, “Nabīdh is unlawful by analogy with wine, for wine was forbidden because it is intoxicating, and that property is present in nabīdh,” this is tantamount to saying, “Every nabīdh is intoxicating, and every intoxicant is unlawful; therefore nabīdh is unlawful.” In this syllogism nabīdh is the subject of the conclusion and hence the minor term; unlawful is the predicate and hence the major term; while intoxicating mediates between subject and predicate—it is the middle term, predicated in the minor premise and serving as the subject in the major.

Again, when you say, “Nabīdh is unlawful by analogy with grape wine, because the cause in the original case is intoxication, and intoxication is present in the branch; therefore the prohibition applies by virtue of its cause,” you are in fact inferring the unlawfulness of nabīdh from intoxication, which is the middle term. Representative analogy simply adds mention of the original case, setting it forth as a model for the branch, because the mind’s apprehension is stronger when it perceives the branch alongside its counterpart than when it merely perceives that both fall under a universal genus. And once proof is established that the attribute is causally effective—