Steve Runge, in his chapter on Information Structure for his upcoming book Discourse Grammar of the Greek New Testament, builds on the principle of Natural Information Flow, the tendency for a discourse to move from what is most known to what is least known (page 133/Chapter 9.2.2; cf. also this comment and following) in conjunction with what Simon Dik’s Theory of Functional Grammar.
Basically*, there are two functional positions in a clause where an author can optionally move an element forward for prominence (Does the “P” in P1 & P2 stand for position or prominent?). When a clause element is moved into one of these positions it violates the natural flow of information. The First position, P1, is for established or topical information in the clause, while the second, P2, is for focused (and non-established) information.
Now that is at the clause level. But when I think about the one hundred pages that I’ve summarized into 24 for my posts on discontinuous syntax in the New Testament, it seems to be only a hop, skip, and a jump to take what has been applied to the clause and apply it to the Greek phrases. This wouldn’t work for English since we use vocal pitch and intonation to express emphasis or focus. But Greek it different. Greek expresses many of these pragmatic functions by means of syntax. This may have something to do with the fact that at some point in its history Greek was tonal (or so scholars believe – I have no idea, myself).
In Part I of my discussion, we saw that the Y1 Hyperbaton structure appears across all phrase types.
Part II examined the claim that the prenominal position in continuous NPs was default for descriptive modifiers** and that the postnominal was default for restrictive modifiers. And finally,
Part III, this morning, made the claim that the initial Y1 position for modifiers in discontinuous phrases was the focused position.
This fits both with the default natural flow of information (NFI) and also withe the prominent P1 & P2 positions:
- Descriptive adjectives tend to be known, established information and thus they default prenominally. This fits the NFI principle.
- Restrictive adjectives tend to be new information, clarifying or restricting the reference of a noun and thus, also in line with NFI principle.
Now I’m less than sure about the P1 position, but it seems rather clear to me that the initial modifier (we’ve been calling it Y1) in discontinuous phrase fits extremely well with P2, the position that specifies focused information. The claim examined in Part III said that in Y1 Hyperbaton, the Y1 modifier received strong focus, evoking and negating a set of alternatives:
ὅλη συγχύννεται Ἰερουσαλήμ (A X N)
all was.in.confusion Jerusalem
“All Jerusalem was in confusion (And not just part!)” Acts 21:31
Look how well this fits with what Steve writes in his grammar,
Placing information in the P2 position represents the choice to take what was already the most important part of the clause (i.e. new and non-established), and to attract even more attention to it by moving it from its default position to a marked one (135).
So, hopefully, the parallels between Dik’s clause level P2 and hyperbaton are as clear on your screen as it is in my head.
This should also be encouraging to those of you who haven’t studied linguistics because mean that separate from a give framework, there’s more consensus on these matter than there are in NT studies (also something that Steve notes in this chapter of his grammar):
Chomskyan generative-transformational theories generally assume that the structure of clauses and phrases are incredibly similar if not identical in languages and DS’s description falls in that camp, suggesting that their pragmatic discussions are parallel with functional conclusions.
Non-Chomskyan generative theories (like LFG & HPSG) while differing in a number of ways from transformation theories use have come to the same conclusions.
Borderline frameworks that seem half functional and half generative such as Van Valin’s Role and Reference Grammar literally uses the same structural representations (nucleus, core, etc.) for both phrases and clauses to make both syntactic parallels and pragmatic ones.
Now I don’t know enough about Functional Grammar to say anything about their phrase level descriptions, but I’d be willing to bet there are similar claims being made there too.