Russell on Descriptions and Denoting [Alberto Torrigiotti]

I am inclined to agree with Russell that there is something curious about sentences which purport to predicate on non-existent entities, namely that certain difficulties arise with regard to their meanings. When I say "the largest even number is an even number", there is a sense in which we want to say this is true by virtue of meaning, but the fact that "the largest even number" denotes nothing (has no reference) precludes us from saying anything meaningful about it; the subject in question does not exist. Secondarily, the logical nature of the questions raised in "On Denoting" suffices to indicate to me that the relation between meaning and denotation is extremely complicated under this technical framework, which is another point on which Russell and I are in agreement.

I am not as convinced as Russell by the supposition that an effort to reduce linguistic elements to logical terms, if done correctly, can yield the sort of profound comprehension of language that Russell seems to assume it can. The best result I can conceive from this approach is an enumeration of true statements, false statements, and statements which are meaningless under this framework, which may be enough for some philosophers of language. To be sure, this approach has its merits, but I can't imagine that it is capable of solving certain kinds of questions we justifiably have about language, or that any conceivable result would give those who presently don't see the necessity or appeal in accepting that the meaning of a statement is just its truth value a compelling reason to take this supposition seriously.

Clarification question: I have a rudimentary understanding of why permitting individual unreal entities membership to the null-class, as Meinong does, seems superfluous. Does Russell just believe that characterizing the null-class as a class of no members is more precise, or does he reject the notion that unreal entities can be members of a class?