But here I should like it to be noted in passing that the more recent Peripatetics, as I at least think, misunderstood the argument of the Ancients by which they strove to prove the existence of God. For, as I find it in the works of a certain Jew, named Rab Chasdai, it reads as follows. If there is an infinite regression of causes, then all things which exist will be things that have been caused. But it cannot pertain to anything that has been caused that it should necessarily exist in virtue of its own nature. Therefore there is in Nature nothing to whose essence it pertains that it should exist necessarily. But this is absurd; and therefore also that. Therefore the force of the argument lies not in the idea that it is impossible for the Infinite actually to exist, or that a regression of causes to infinity is impossible, but only in the impossibility of supposing that things which do not exist necessarily in virtue of their own nature, are not determined to existence by something which does exist necessarily in virtue of its own nature, and which is a Cause, not an Effect.
This is an interesting interpretation of infinite regress arguments. It means that I'll certainly have to look at what Hasdai Crescas (one of the great Jewish thinkers of the Middle Ages) actually says on the subject, to see how much of this is Rab Hasdai and how much is Spinoza's interpretation of him. Spinoza is right that the usual reasons given for the impossibility of infinite regress are (1) that infinite regression of the given sort of cause involves a contradiction; and (2) that it is impossible for an actually infinite multitude, of the sort infinite regress requires, to exist. Without having Crescas before me, however, I rather suspect that Spinoza may have misread him; the only reason I have for thinking so is that people have occasionally misinterpreted Aquinas this way -- any elaboration of (1) will look superficially like the other. But if Crescas has this interpretation, it's of interest, since it's not the usual interpretation of this type of argument.