The expression ∀x(ϕx → ψx) is supposed to mean that, in Russell's parlance, ϕx → ψx is true "for all values of x". However, what are those values that Russell is referring to?
At some point in a now distant past, all mathematicians would have probably happily subscribed to the idea that if x is a number, then x² is positive or null. Russell himself would have said that "x is a number" implies "x ≥ 0" for all values of x. Yet, after mathematicians decided that it was acceptable to say that there is a number i such that i² = -1, the expression ∀x("x is a number" → x ≥ 0) is no longer true.
I suppose that this sort of problem is solved by the notion of domain, whereby one can specify in advance that x belongs to some restrictive set of values, for example, in this case, ℕ. If the domain is ℕ, then "x is a number" implies "x ≥ 0" for all values of x if only because for all values of x now means "for all values of x in the domain", which means in this case x ∈ ℕ.
Thank you to correct me if I am wrong there.
Still, I couldn't find where Russell would have introduced the notion of domain as it is now defined in all mathematics textbooks. So, my question is:
Did Russell had something like the notion of domain in the sense defined now by mathematics textbooks?
Thank you for any scholarly references.
forall n in N, P
is just short-hand forforall n, n in N -> P
. Namely, quantification in the underlying language (e.g. FOL) is not restricted: quantified variables simply range over all possible individuals (individual constants).