No, I don’t have any insightful new conclusions on this, but I did run across a very interesting instance that warrants further study and research.

First, note that most grammars portray πλην as a strengthened form of αλλα, and portray αλλα as a “strong adversative” in comparison to δε, which they class as “adversative” (at least in one of the typically myriad senses). I’m not happy with that. I get the idea of a cline here, but it just seems sloppy. What, there’s an “adversative”, a “strong adversative” and a “stronger adversative”? How does that help? How does that not simply mistake contextual cues for lexical/grammatical function?

I think the insights from Stephen Levinsohn, Steve Runge and Discourse Grammar help a bit more. That is, δε indicates development; αλλα is a marker of contrast and indicates correction or replacement (again, see my αλλα paper).

But what does πλην do, then? Muddying up the works is the notion that πλην can function as a conjunction (most common) or as a preposition (Ac 15.28, cf. Robertson XIII.VIII(ai), p. 646), or as a simple adverb (Ac 20.23; Php 1.18, again cf. Robertson).

One approach is to see how the same author uses these two function words. Last night, while working through the text of Mt 26, I came upon the following that screams out to me for further examination/research:

αὶ προελθὼν μικρὸν ἔπεσεν ἐπὶ πρόσωπον αὐτοῦ προσευχόμενος καὶ λέγων· πάτερ μου, εἰ δυνατόν ἐστιν, παρελθάτω ἀπʼ ἐμοῦ τὸ ποτήριον τοῦτο· πλὴν οὐχ ὡς ἐγὼ θέλω ἀλλʼ ὡς σύ. (Mt 26.39, NA27)

and going forward a bit, he fell upon his face praying and saying, "My Father, if it is possible, remove from me this cup; πλην not as I desire αλλʼ as you [desire]." (Mt 26.39, my own)

The synoptic parallels are interesting (thanks for pointing them out, Steve), particularly depending on how you view their composition histories. Me? I think ‘Q’ is a load of hooey, and think considering Mark as the first written and also as a source for both Mt and Luke makes sense. So here’s what Mark and Luke (respectively) have:

καὶ ἔλεγεν· ἀββα ὁ πατήρ, πάντα δυνατά σοι· παρένεγκε τὸ ποτήριον τοῦτο ἀπʼ ἐμοῦ· ἀλλʼ οὐ τί ἐγὼ θέλω ἀλλὰ τί σύ. (Mk 14.36, NA27)

and he said: “Abba Father, all things are possible with you. Take this cup away from me, αλλʼ not what I desire, αλλα what you [desire].” (Mk 14.36, my own)

λέγων· πάτερ, εἰ βούλει παρένεγκε τοῦτο τὸ ποτήριον ἀπʼ ἐμοῦ· πλὴν μὴ τὸ θέλημά μου ἀλλὰ τὸ σὸν γινέσθω. (Lu 22.42 NA27)

saying, “Father, if you wish, take this cup away from me, πλην not my desire αλλα your [desire] be done.” (Lu 22.42, my own)

So Mark uses αλλα … αλλα while Luke (along with Matthew) use πλην … αλλα. That’s very interesting too; it means that examining the higher-level discourse of all three gospel writers in this instance could shed some light on why Matthew and Luke both prefer πλην … αλλα.

Might have to consider digging into this for the ETS NW Regional meeting. It’s in February, and it’s only in Tacoma.

Update: For those keeping score at home, note that Margaret Thrall addresses this very issue in her book Greek Particles in the New Testament. Whether proper or not, you can grab a copy of Thrall from archive.org (search "Texts" for "Margaret Thrall"). The gist, apparently, is that she sees Mark as the strongest formulation, with πλην in Mt & Luke functioning conditionally in this instance. Her discussion runs from pp. 67-70. Davies & Allison (ICC Matthew vol. 3) agree with Thrall; Luz (Hermeneia Matthew 3) disagrees saying in a note "Here the adversative sense dominates". I'm inclined to go against Thrall as well but will actually read the section first (of course).