I haven't made any study of Sharp's rule, and I know there is a current ongoing debate about its empirical basis in Greek scholarship. I can however make some more general linguistic points.
When one talks about "rules" one must distinguish between prescriptive and descriptive rules.
1. Prescriptive rules are programmatic and place demands on linguistic usage, whereas
2. descriptive rules generalize observations on how a specific language works.
There are also two main kinds of descriptive rules:
1.I-language grammar, which refers to the inferred internal, intuitive cognitive rules that generates a speaker's linguistic output, and
2. E-language grammar, which refers to the external, community-based level of grammar that embraces the heterogeneity of a multiplicity of idiolects.
Sharp's rule is:
A.a generalization based on a limited corpus of Koine Greek, the NT.
B. It is an abstraction based on actual stylistic usage,
C. it is an E-language descriptive rule.
It is not a prescriptive rule and is only reflective of possible I-language rules; it does not itself rule out the possibility that there were I-language grammars that did not include such a rule, as Koine Greek was heterogenous and contained a lot of grammatical variation.
So it is important to first understand that rules devised by Greek scholars are abstractions that do not necessarily correspond to actual mental constructs in the heads of ancient Greek speakers that governed how they wrote and spoke. Unless it is a core grammatical rule that must be respected categorically for the language to work (violations of which would produce sentences that a given speaker would reliably regard as ungrammatical if not incomprehensible), it is a probabilistic generalization of a stylistic tendency of usage and thus does not necessarily demand its observance.
So I think it overstates things to say that such a rule, even if 100% valid, "proves" that a text must be read in a manner consistent with it. And Sharp's rule concerns a form of nominal coordination (TSKS constructions), which strikes me as a less "core", more fluid area in a language's grammar.
The flip side to this is that such rules are generalizations of actual linguistic usage, so counterexamples do not necessarily invalidate their descriptive utility.
It isn't a matter of disproving the rule by finding a counterexample or two in the corpus, unless of course the corpus is so tiny that such counterexamples have a major effect on determining whether there is a grammatical tendency or not.
So even if the rule isn't categorical, it does have some weight in assessing a given text like Titus 2:13, and may offer a strong argument for preferring one rendering over another.
That is a different matter than saying that it "proves" that Translations X,Y,Z are valid or invalid.
The NT is a quite small corpus, but the validity of the rule may be tested further by checking other closely related corpora to see if it holds there as well.
The LXX is an important corpus because it is a direct source for much of the NT; at the same time it is quite distinct because it has a level of interference not found in much of the NT and antedates the writing of the NT by at least a century (while at the same time corrected and revised for hundreds of years later).
A more closely related corpus is that of the apostolic fathers, which overlaps the time period of the NT's composition and includes writings that were sometimes incorporated into the NT before the canon was finalized.
I am not clear on the details but my memory is that Sharp's rule holds up quite well in these other corpora.
That isn't to say that there aren't possible counterexamples or ambiguous cases, but by and large it holds up in an overwhelming majority of the examples.
I am sure the current discussion of the rule in scholarship concerns just how well the rule does hold up, i.e. how robust a generalization it is. If it is, then it offers a strong argument for handling Titus 2:13 in a manner that respects it. But I am somewhat doubtful a constructed "rule" based on usage can demand a given rendering, or "prove" a rendering is correct beyond doubt.
I have personally regarded passages like Titus 2:13, 2 Peter 1:1, Acts 20:28, etc. as ambiguous, and I have wondered if perhaps they are intentionally ambiguous and represent one stage in the development of a more explicit expression of higher christology.
Terry says: (I actually get claustrophobic panic attacks when confronted with dense paragraphs of complex sentences. The relief comes if I break it up into constituencies of separate thoughts even if it is artificially wrought.)
MY QUESTION: Do such scriptures reflect orally transmitted ideas being kicked around over time and detached from the actual persons originating them?
Or, do they embody the actual thoughts of the writer (who is thinking them through)?
Would the writer (or the originator) of these passages presume to know with Authority or merely be speaking opinion with a kind of self-assured certainty?
And finally, HOW WOULD WE EVER BE ABLE TO TELL??
The reason I ask is that I wonder at the eager acceptance of readers detached and separated by two thousand years from these oft-translated passages as to their OWN confidence that actual "information" is being transmitted intact!