I've been trying to understand in a more formal way what a set actually is, but I have some questions. According to the axiom of regularity for every non-empty set A there exists an element in the set that's disjoint from A. That would mean that such element is also a set, right?

I read here: Axiom of Regularity , that in axiomatic set theory everything is a set, I understand that natural numbers are constructed from the empty set, integers are constructed from the naturals, rationals from the integers, and reals from the rationals. I can see how every element in such sets is also a set. But, for example, in the set of all the letters of the alphabet, or the sample space of an experiment when the possible results are not numbers, or the set of my classmates; it's not clear to me how their elements are also sets. So, are they really sets? Is every element in a set also a set?

Thank you

  • 1
    $\begingroup$ If you're not doing axiomatic set theory, but instead topology, or analysis, or algebra, or something, then it's safe to think that the most basic elements / points are not sets IMO. $\endgroup$
    – Arthur
    Feb 8 '17 at 7:30
  • 2
    $\begingroup$ See urelements : we nay have set theories with them, i.e. objects that are not sets but that can be elements of sets, like e.g. natural numbers, and we may have set theories without urelements, i.e. every object is a set. $\endgroup$ Feb 8 '17 at 7:33
  • 1
    $\begingroup$ Original (1908) axiomatization of set theory by Zermelo was with urelements. $\endgroup$ Feb 8 '17 at 7:34

In the context of axiomatic set theory, we (usually) only allow sets whose elements are sets. So in that context, there is no such thing as "the set of all the letters of the alphabet", or your other examples. However, that doesn't mean you can't talk about these concepts in axiomatic set theory--you just have to encode them in sets. For instance, instead of talking about the letters of the alphabet, you could talk about the set of natural numbers from $1$ to $26$, with the understanding that $1$ secretly stands for A and $2$ secretly stands for B and so on. This isn't really any different from the idea of "constructing" numbers as sets which you said you were familiar with: we have an intuitive idea of numbers, and we find a way to talk about them using only sets whose elements are all sets. Similarly, we can find such a way to talk about letters, or classmates, or whatever else you want to talk about that is not intuitively a set.


That depends on which axioms or system you're using. You could of course make up a system which allows (distinct) atomic objects (they're called ur-elements) that can be elements of sets, this is how ZF set theory originally started.

However in (modern) ZF set theory the axiom of extensionality basically prohibits anything that's not a set. Things are considered equal if they have the same elements and since anything not being a set does not have any elements they would be considered equal to the empty set.

As you pointed out one constructs the natural numbers and so on by constructing set. However one normally does not use those properties of them (being sets), you almost never see things like $1 \cup 2$ or $0 \in 1$. One should note that the way natural numbers are constructed is not standardized, that is there is different ways to achieve the same (standardized) properties of the numbers (which makes using numbers as they are sets being non-standard and have no universally accepted meaning).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.