Page 1 of 1

New dissertation on Greek Word Order (Recht, UC Berkeley)

Posted: May 23rd, 2015, 7:39 am
by Stephen Carlson
I just came across a new dissertation on Greek word order with special emphasis on verb-initial clauses: Here is the abstract:
Recht 2015 dissertation abstract wrote:Word order in Ancient Greek, a ‘free word order’ or discourse-configurational language, depends largely on pragmatic and information-structural factors, but the precise nature of these factors is still a matter of some controversy (Dik 1995, Matić 2003). In this dissertation, I examine the set of constructions in which a verb appears in first position in its clause, and consider the conditions under which such constructions appear and the roles they play in structuring Greek discourse. I distinguish between topical and focal initial verbs, and show that the former class (which are the main concern of the study) in fact occur as part of larger units definable in terms of both prosody and pragmatics. The function of such units, I argue, is to mark specific kinds of transitions between the implicit questions that structure discourse (Questions Under Discussion [QUDs], Roberts 1996). I describe and categorize the types of QUD transitions marked by verb-initial units in a corpus of five fifth-and fourth-century Greek prose authors, and relate these to transitions marked by other classes of constructions, including a newly identified contrastive-topic construction. My account improves on preceding models by unifying a number of phenomena previously treated as disparate. It also represents the first large-scale application of the QUD model to real discourse.
I haven't read the thesis yet and so cannot vouch for it at this time, but I've come to conclusion that prosody and pragmatics are critical for understanding Greek word order.

Re: New dissertation on Greek Word Order (Recht, UC Berkeley

Posted: May 23rd, 2015, 1:37 pm
by Jonathan Robie
I'm still in the first part of this, but this guy writes well and thinks clearly. I'd like to drop a few things into this thread as I read for discussion.

I thought this was interesting:
Surely every scholar is acquainted with the stories of Plato’s passion for taking pains, especially that of the tablet which they say was found after his death, with the beginning of the Republic (“I went down yesterday to the Piraeus together with Glaucon the son of Ariston”) arranged in elaborately varying orders.

Dionysius of Halicarnassus, On Literary Composition 25.209
That's a good illustration that writers were conscious of word order, and a reminder that they had erasable wax tablets and used them.

I'm guessing that Plato did this more or less the same way modern English writers do this, writing out various versions of the sentence with intuition as a guide. I'm guessing he wasn't asking himself whether a given order implied topic vs. focus vs. emphasis, and that he wasn't thinking in terms of SVO vs. VSO etc. I suspect this is what Stephen Carlson means when he says he thinks that prosody and pragmatics are behind this.

Re: New dissertation on Greek Word Order (Recht, UC Berkeley

Posted: May 23rd, 2015, 1:48 pm
by Jonathan Robie
Here's the author's assessment of our current state of knowledge with respect to classical Greek word order, and the kind of language that Greek is:
Ancient Greek is a ‘free word order’ language, or more precisely a discourse-configurational language (Hale 1983, Kiss 1994): that is, a language in which
the order of words in a sentence is determined – at least in the case of clause-level constituents, though to some extent in lower ones too – not by their syntactic roles but by the pragmatic functions that they play in the discourse context. The question of how to understand these discourse-pragmatic factors and their effects on Greek sentence structures is almost as old as the study of Greek grammar itself, though recent decades have seen considerable progress in this field thanks to the application of updated theoretical concepts in pragmatics and information structure. Still, this work has only touched the surface of the multifarious and sometimes baffling array of word-order phenomena that even the simplest Greek prose texts display, and many important constructions have received little attention.
This dissertation is limited to one particular construction, the verb-initial clause.

Re: New dissertation on Greek Word Order (Recht, UC Berkeley

Posted: May 23rd, 2015, 2:32 pm
by Stirling Bartholomew
Someone trying to figure out what is going on in this thesis, might get as far as page 60 before getting anywhere. Here are two sections which might help.

Preposed Verb-initial Units (PUV) are not defined by syntax:

page 48
1 Prosodic segmentation in Greek
A fact that has often gone unremarked in previous analyses of clause-initial verbs
in Greek is that many such verbs are fronted along with additional material: that
is, initial verbs can be followed by one or more other constituents or subconstituents
with which they appear to form a unit. Such units are definable in both
informational and prosodic terms. Understanding the functions of initial verbs, as
I hope to show in this study, requires an understanding both of these additional
elements and of the function of these units as a whole.
I will refer to such initial units—both those consisting of a verb by itself and
those consisting of a verb followed by one or more other elements—as ‘preposed
verb-initial units’ or PVUs.
page 60
3 The status of non-verbal elements in PVUs
The inventory just given has shown that the conditions licensing inclusion of nonverbal
elements in a PVU cannot be syntactic, as all kinds of syntactic material
can appear in PVUs. Rather, these conditions are discourse-pragmatic. Specifically,
I argue that the inclusion of non-verbal elements in a PVU is licensed by the
following discourse-pragmatic criterion:
The referents of non-verbal elements in a PVU must either (a) be
discourse-active or discourse-inferable, or (b) form part of an introduced
topic together with the verb. Stated otherwise, they must be
part of the common ground, either (a) actually, or (b) by requested
accommodation.
This criterion holds for all PVUs in my corpus. Before proceeding to the discussion
of examples, I should say that the distinction between conditions (a) and
(b) is intended as practically useful rather than theoretically watertight, since
even newly introduced topics can often be regarded as discourse-inferable in some
way. (Phrased in terms of the Question Under Discussion framework, a new question
usually stands in some definable relation to a previously active question.).
As was pointed out in the previous chapter, introducing a new topic is equivalent
as a discourse move to a request that it be accommodated as part of the
common ground, so the difference between ‘new topic’ and ‘discourse-inferable
topic’ comes down to a question of how much accommodation is being requested,
so to speak. In terms of the QUD model, both types of elements can be described
as expressing part of the QUD being introduced – which is the main function of
PVUs generally, as the next chapter argues – rather than part of its answer. I
will have more to say below about the broader generalization underlying the two
types.

Re: New dissertation on Greek Word Order (Recht, UC Berkeley

Posted: May 23rd, 2015, 2:49 pm
by Thomas Dolhanty
Jonathan Robie wrote:That's a good illustration that writers were conscious of word order, and a reminder that they had erasable wax tablets and used them.

I'm guessing that Plato did this more or less the same way modern English writers do this, writing out various versions of the sentence with intuition as a guide. I'm guessing he wasn't asking himself whether a given order implied topic vs. focus vs. emphasis, and that he wasn't thinking in terms of SVO vs. VSO etc. I suspect this is what Stephen Carlson means when he says he thinks that prosody and pragmatics are behind this.
I thought that one tool which would help me, and hopefully others, would be a little program which would take a GNT sentence / clause / phrase as input, scramble the words, and then allow the user to try to reassemble the words in the correct (or perhaps "a" correct) order. So, one would upload a series of verses / sentences / clauses, etc., and then select whichever verse (eg.) one wanted to consider at the time, hit the "scramble" button, and then move the words around to 'rebuild' the sentence.

I already have a crude working version in Excel, and will work on it for a bit, and then share it if anyone is interested. Eventually, I would like to refine it so that one can select any GNT verse from an XML file of the entire GNT.

Re: New dissertation on Greek Word Order (Recht, UC Berkeley

Posted: May 23rd, 2015, 6:47 pm
by Stephen Carlson
I'm a third of the way through the thesis, but I'm gratified to see that he's interacting with the same people whose work on Greek word order (H. Dik, Matic, Goldstein, Scheppers) whom I've found to be helpful.

A minor demerit is that he didn't realize that Katalin É. Kiss surname begins with the É. She's even written up in the Chicago Manual of Style.

Re: New dissertation on Greek Word Order (Recht, UC Berkeley

Posted: May 24th, 2015, 1:58 am
by Stephen Carlson
Now that I've a chance to read over the thesis, my initial impression is somewhat lukewarm. I'm not sure Craige Robert's QUD (Question under Discussion) approach is as helpful as he think it is; it seems to just push the obscurity in the text back another level. But it's nice to see her approach get some play (Recht passes a note from her that his is the first large-scale study using it), and I would have to digest it some more before I know what to do with it.

I am also skeptical that his PVUs (preposed verbal-initial units) are a coherent category. I would tend to consider these units not as preposed but what's left after dislocating material to the right in a tail or elaboration. His brief consideration of this possibility on p. 50 and rejection thereof on the grounds of lacking syntactic constituency seems weak to me, considering that he has no problem with PVU's lacking syntactic constituency.

Nevertheless, I am happy to see him take seriously the possibility of post-verbal focus, which I have found challenging. I will illustrate these with examples of my own from Galatians 2.

One example of his contrastive topic construction (adding | to indicate a prosodic unit boundary his analysis presumes) is:
Gal 2:20 wrote:ζῶ δὲ οὐκέτι | ἐγώ,
ζῇ δὲ ἐν ἐμοὶ | Χριστός·
Another example would be his additive counter-presuppositional focus:
Gal 2:13 wrote:καὶ συνυπεκρίθησαν αὐτῷ | καὶ οἱ λοιποὶ Ἰουδαῖοι, ...