Currently, I'm rereading Steven Runge's Discourse Grammar of the Greek New Testament: A Practical Introduction for Teaching and Exegesis. It isn't a full-fledge discourse grammar but contains elements of discourse features when approaching the NT text.
Chapter 2: Connecting Propositions (sample PDF) is the worth the price of the book, 3-times over. In this chapter, he speaks of pragmatic definitions of conjunctions in modern grammars. The problem is, grammars utilize english and other functional categories to describe conjunctions. To provide an example, Daniel Wallace's Greek Grammar Beyond the Basics is a case in point. Consider the amount of overlap between the following conjunctions (Wallace, Greek Grammar, 761):
Logical Functions
A. Ascensive: even...καί, δέ, and μηδέ
B. Connective: and, also... καί and δέ
C. Contrastive (adversative): but, rather, however... ἀλλά, πλήν, sometimes καί and δέ
D. Correlative: μέν...δέ (on the one hand...on the other hand); καί...καί (both...and)
E. Disjunctive (alternative): or...ἤ
F. Emphatic: certainly, indeed...ἀλλά (certainly), οὐ μή (certainly not or by no means), οὖν (certainly); true emphatic conjunctions include γε, δή, μενοῦνγε, μέντοι, ναί, and νή
G. Explanatory: for, you see, or that is, namely...γάρ, δέ, εἰ (after verbs of emotion), andκαί
H. Inferential: therefore...ἄρα, γάρ, διό, διότι, οὖν, πλήν, τοιγαροῦν, τοινῦν, and ὥστε
I. Transitional: now, then...οὖν and especially δέ
Wallace provides functional categories based upon english grammatical concepts. One is really left with very little distinction, especially between, καί and δέ. They both appear in the similar categories; therefore, one is left to assume they have no pragmatic difference.
Runge seeks to revert to semantic categories instead of pragmatic distinctions. Καί can have connective and adversative functions, but its semantic value, according to Runge, "links items of equal status" (Runge, 26).
Δέ traditionally has been understood as a disjunctive particle, but it too can have connective, transitional, and contrastive functions. It is also used with μέν...δέ clauses too, which do not necessarily function as a connective or contrastive. Instead of traditional grammatical definitions for δέ, Runge attempts to define δέ as a "developmental marker" (Runge, 29). That is, it is not necessarily developing the temporal development of the argument, but the logical development (Runge, 36).
Therefore, καί and δέ can function very similarly but they have two distinct semantic meaning. Runge does not use Matt 1:2–3 as an example. I will provide it here to demonstrate the differences between καί ("linking items together of equal status") and δέ (developing the argument).
- 2 Ἀβραὰμ ἐγέννησεν τὸν Ἰσαάκ,
- Ἰσαὰκ δὲ ἐγέννησεν τὸν Ἰακώβ,
- Ἰακὼβ δὲ ἐγέννησεν τὸν Ἰούδαν καὶ τοὺς ἀδελφοὺς αὐτοῦ,
- 3 Ἰούδας δὲ ἐγέννησεν τὸν Φάρες καὶ τὸν Ζάρα ἐκ τῆς Θαμάρ,
- Φάρες δὲ ἐγέννησεν τὸν Ἑσρώμ,
- Ἑσρὼμ δὲ ἐγέννησεν τὸν Ἀράμ,
Translation:
- Abram begot Isaac
- Isaac begot Jacob
- Jacob begot Judah and his brothers
- Judah begot Perez and Zerah from Tamar
- Perez begot Hezron
- Hezron begot Aram
Noting each bold word, δέ begins a new idea and therefore develops the flow of the argument. Each italic word fits between two δέ clauses, linking two closely related ideas. In this case, Runge's paradigm holds true. Furthermore, the rest of Matt 1 continues this idea of καί serving as closely "linking items of equal status" whereas δέ serves as a "developmental marker" highlighting a new development (either disjunctive or continuative) in the storyline.
Runge is on to something, especially his understanding of conjunctions. He currently blogs at NT Discourse and serves as the Resident Scholar at Logos Bible Software.