In grammar and semantics, linguists talk about "prototype theory
." Your "exceptions" are a great example of non-prototypicality. Sharps rule describes two coordinated singular, common, nouns sharing an article. That's the basic morpho-syntactic conception. When you add semantics to it, things change. The prototypical instance of the rule is always going to involve two common nouns with no inherent relationship to each other.
Prototype theory sounds like an excellent way to go about understanding the Granville-Sharp construction--I'm with you there--but that's not the approach Wallace used. Rather, his approach was a lot more categorical, trying to determine the exact, necessary and sufficient conditions in which the rule held.
I must admit that I found such a categorical approach quite frustrating. Many of the qualifications and exceptions seem poorly motivated except to fit a fairly small corpus. Frankly, that proposed Trinitarian exception was the last straw for me on his categorical approach, especially since he seemed very keen on applying a suitably refined Granville-Sharp rule to Tit 2:13 and 2 Pet 1:1 in order to affirm a major element of Trinitarian theology.
There's got to be a better way for conceptualizing this construction, and perhaps prototype theory is the way to go. I note that the Classical Greek grammars don't seem to recognize the specific construction. Rather, they seem content to address a more general contraction with a conjunction of two or more nouns governed by a single article, without the specific qualifications and exceptions that Wallace identifies.
For example, Stéphanie Bakker, The Noun Phrase in Ancient Greek
(ASCP 15; Leiden: Brill, 2009), 177 states:
In a (very small) number of cases of coordinations of identifiable entities, however, one article serves to express the identifiability of all the entities. The effect of the omission of the article with the second, third, etc. noun in the coordination is that the different entities are depicted as one whole.
Her footnote (omitted above) cites Smyth § 1143 who states:
1143. A single article, used with the first of two or more nouns connected by and, produces the effect of a single notion: οἱ στρατηγοὶ καὶ λοχᾱγοί the generals and captains (the commanding officers) X. A. 2. 2. 8, τὰς μεγίστᾱς καὶ ἐλαχίστᾱς ναῦς the largest and the smallest ships (the whole fleet) T. 1. 10, ἡ τῶν πολλῶν διαβολή τε καὶ φθόνος the calumniations and envy of the multitude P. A. 28 a. Rarely when the substantives are of different genders: περὶ τὰς ἑαυτῶν ψῡχὰς καὶ σώματα concerning their own lives and persons X. A. 3. 2. 20.
So, I wonder if defining such a narrow subcategory of coordinations of nouns governed by a single article is all that useful, especially compared to a prototype approach.
Stephen C. Carlson, Ph.D.