Response from Daniel J. Velleman on October 12, 2005

Russell's Paradox is a problem for set theory--or at least it was when Russell discovered it. The most popular modern approach to set theory is based on the axioms developed by Zermelo and Frankel, and the Zermelo-Frankel (ZF) axioms are formulated to avoid the paradox. So Russell's Paradox is not a problem for modern set theory.

The reason paradoxes in set theory are considered to be such a serious matter is that most mathematicians regard set theory as the foundation of all of mathematics. Virtually all mathematical statements can be formulated in the language of set theory, and all mathematical theorems--including your example 2+2=4--can be proven from the ZF axioms.

But you ask about our "confidence" that 2+2=4 is true. I don't think anyone's confidence in 2+2=4 is based on the fact that it is provable in ZF set theory, even though ZF is regarded as the foundation of mathematics. It's hard to imagine anyone having serious doubts about whether or not 2+2=4, and having those doubts relieved by seeing the proof in ZF. So if a new paradox were discovered that showed that the ZF axioms were flawed, I don't think anyone's confidence in 2+2=4 would be shaken. We'd just try to fix the flaw in the axioms in a way that would allow us to still develop mathematics.