What does 1a demonstrate about expressions?

Hm,

I think it means 1 multiplied by a random number, or variable in this case.

After all, the numerals in (1a-b) and the numerals in (3a-b) are not plausibly homonyms ... Co-referringexpressions are often not substitutable, so (5) does no more than further demonstrate the inadequacy of the naïve reference principle.