Can a pure class be an atom? i.e., can there be a pure class that is a member of another class?
This question is from Maschovakis' Notes on set theory (2nd edition).
He says that "We assume at the outset that there is a domain or universe $\cal W$ of objects, some of which are sets, and certain definite conditions and operations on $\cal W$. We call the objects in $\cal W$ which are not sets atoms, but we do not require that any atoms exist. Definite conditions and operations are neither sets nor atoms."
"We will profess that for every unary, definite condition $P$ there exists a class $A=\{x|P(x)\}$ such that for every object $x$, $$x\in A\iff P(x)$$."
In exercise 3.23, he asks to prove: For every class $A$, $$\text {$A$ is a set $\iff$ for some class $B$, $A\in B$ $\iff$ for some set $X$, $A\subset X$. } $$ where inclusion among classes is defined as they were classes.
The author provides solution to this exercise, which for $(2)\implies (3)$ is as follows: We will assume that $A$ is a class. If $A$ belongs to some class $B$, then it must be an object (atom or set), since only objects are put into classes by definition of class; and since $A$ is a class by the hypothesis, it is not an atom, and so it must be a set. Since $A\subset A$, $A$ is a subset of a set.
What I can't understand is that how could we conclude that a class can't be an atom.