Weird. I was editing along on this discourse handbook on the book of James when I noticed the phrase δὲ ἐν.... It occurred to me, why isn't the ε elided to give δʼ instead? For that matter, how often does δέ get elided in the GNT? I don't recall it happening that often. It's fairly common in Classical Greek (it occurs over 3000 times in LSJ alone), so I just didn't think about it much until today.
So, I did a search for δʼ in the GNT. Ready? In the SBLGNT it occurs 28 times (including the apparatus) and in NA27 it occurs 24 times (also including the apparatus). Not very common. But get this! In both editions, all but one of the occurrences is δʼ ἄν! The only other occurrence is in Matthew 27:44 where it occurs as δʼ αὐτό.
Fascinating! Why the lack of elision in the GNT? I need to look at the apparatus in NA27, too. The witnesses should be interesting...
By the way, in Classical δὲ ἐν does elide, so that isn't a problem. I don't have time to chase it right now, but I wonder if this is a synchronic occurrence? Do the Apostolic Fathers elide δέ? What about Josephus and Philo? Xenophon?
[Update]I posted to the Classics-L e-mail list and Claude Pavur suggested I check the Septuagint (LXX). Here's what I found: 60 occurrences (including the apparatus), of which the majority are δʼ ἄν, but the more "regular" elisions occur also. It doesn't seem to matter whether the books were composed originally in Greek or not.
Here's an strange one, in Wisdom 16:23: τοῦτο πάλιν δʼ, ἵνα τραφῶσιν δίκαιοι
I doubt the wisdom (pun intended) of that text critical choice—elision across a clause boundary? Seems strange. The text critical note: παλιν δ B†] δ > S, δε παλιν A
In English, that means the B (Vaticanus) has the reading, S (Sinaiticus drops the δ, and Alexandrinus puts the δε in front of παλιν.
More, probably useless, information...nothing to see here, carry on? Or, these are the droids you're looking for? : )