[b-greek] Re: Constituent order

From: Nicholas A Bailey (nicholas_bailey@sil.org)
Date: Fri Nov 23 2001 - 10:24:44 EST

> After seeing Clay cite Levinsohn's book on discourse features of NT
> Greek several times,** I decided to have my local library get it for
> me through interlibrary loan, which led me to buy the book for
> myself (dating before marriage is a good rule for books).
> I'm still working through the book, but I wondered if others could
> comment on his approach to constituent order. Is his approach
> generally accepted by linguists, or is he building his whole system
> on questional assumptions? How do other approaches differ?
> For those who don't have his book, he essentially says that if a
> sentence begins with a nonverbal constituent, it will either be a
> "point of departure" or "in focus." (p. 45). A point of departure
> "provides a starting point for the communication" and "cohesively
> anchors the subsequent clause(s) to something which is already in
> the context (i.e., to something accessible in the hearer's mental
> representation.)." (p. 8). "The focus of an utterance is that part
> which is intended to make the most important... change in the
> hearer's mental representation." (p. 32 n. 8).
> I enjoy discussing books that I'm reading with people, especially
> people who know a lot more than I do. Any discussion would be
> welcome.
> Jonathan Boyd
> Huxley, IA
> **Stephen H. Levinsohn, _Discourse Features of New Testament
> Greek_, 2d ed. (Dallas: SIL International, 2000).

November 23, 2001

Having seen this query about Stephen Levinsohn’s work (who is a colleague
of mine) I couldn’t resist making a some comments.

Simon Crisp of UBS has offered some comments and evaluation of Levinsohn’s
work in the UBS newsletter Tic Talk 37, 1997. Check out this site:


Crisp brings up especially comments by Stanley Porter, but it is clear to
me (and I recall Crisp implying) that Porter somewhat misunderstands
Levinsohn’s work. Levinsohn is interested in discourse analysis of NT
Greek, but part of the problem is that to understand discourse properly,
(which involves the use of utterances or sentences in context) one has to
understand especially the internal structure of sentences themselves. You
have to understand how the nuts and bolts fit together if you want to
understand how everything in the machine functions.

Porter also felt that Levinsohn’s work is not mainstream, at least among
Greek scholars. Perhaps that is changing, and maybe even Porter’s opinion,
but in any case, Levinsohn’s linguistic perspective, while somewhat
eclectic, is fairly representative of a lot of work done in both
Typological linguistics and Functional linguistics (including Functional
Grammar, and earlier the Prague school, see work by Talmy Givón, Simon
Dik, Firbas, Benes, and also some discourse analysis linguists like
Wallace Chafe, Knud Lambrecht, Ellen Prince).

What some discourse linguists call “information structure” concerns the
“clues” (formal linguistic expressions) in utterances (sentences) that
help a hearer (reader) know how these sentences relate to discourse. Two
fundamental categories of information structure are 1) what is part of the
common ground (“presupposition”) between the speaker and hearer, and 2)
what is the new idea that the speaker wants the hearer to realize by
making the utterance (assertion, or what is “asserted”). Most or perhaps
all communication relies on a relationship between these two cognitive
categories, by relating a new idea to some kind of background or knowledge
presupposition. The terms topic/point of departure and focus, which
Levinsohn uses, are subcategories of presupposition and assertion. Here
are some general correlations:

Presupposition : topic/point of departure (what is given, old or to be
taken for granted)

Assertion : focus (new information)

(See especially Knud Lambrecht of the University of Texas at Austin, who
is a leader in the field of information structure analysis. Knud Lambrecht
1994. Information Structure and Sentence form. Cambridge: Cambridge
University Press.)

You will notice that Levinsohn begins his discussion (page 7) by referring
to sentence articulations. These are the three most basic kinds of
information structure “combos”, or ways of “packaging” utterances in terms
of presupposition and assertion. Levinsohn mentions topic-comment,
focus-presupposition and presentational articulations. (See also Lambrecht

Many languages, especially European ones, use intonation to (among other
things) at least partially indicate information structure, but many
languages (not so much English) also use word order and particles to
indicate information structure. Since we cannot say anything about the
intonation of a dead language like Greek, we are left to sorting out the
remainder of the clues, which show up in the word order, certain
particles, and syntactic forms. So much of Levinsohn’s work is devoted to
exactly this messy job of sorting all these factors out. And the
frustrating thing is that we are probably missing one of the most
important clues, intonation, assuming that Greek intonation did indeed
have one of these functions.

While Levinsohn’s work is ongoing, and will inevitably involve revision,
one of his real contributions is how he has shown that the sentence
initial position is used to indicate both point of departure/topic and
also focus. Other positions of the sentence though also seem to regularly
indicate focus, depending on the kinds of syntactic situations they come
in. And focus itself is subdivided into at least run-of-the-mill focus and
more emphatic varieties that include at least “contrastive focus”.

To illustrate the problem a bit better, consider how commentaries and
Greek grammars may claim that such-and-such a word is “emphasized” in the
Greek. This is rather frustrating because we are left wondering in what
way it is emphasized. To explain this problem, I will try to illustrate
here how English uses intonational stress to indicate both point of
departure (or topic change) and focus of assertion. Consider the following
sentences where CAPITAL letters indicate intonational stress, and where
for each is given a general context indicated in (…). (These examples are
modified ones from Lambrecht 1994).

1. (A: What did the children do next?)
B: The Children/They went to SCHOOL.
information structure: topic-comment

2. (A: So what did everyone do today? the children, grandmother, parents
information structure: 2 clauses with a point of departure which functions
as topic plus comment

3. (A: Who went to school? (the speaker is asking real question.)
B: The CHILDREN (went to school.)
information structure: argument focus where an “open” presupposition like
“X went to school” is taken for granted but where there is a request to
“fill in the blank”, indicated by “who?”

4. (A: So I heard that the parents went to school. (From B’s perspective
this belief is false.))
B1: (No,) The CHILDREN went to school.
B2: (No,) It was the CHILDREN who went to school.
information structure: argument focus where a part of the presupposition
like “the parents went to school” is taken for granted but needs
correction, i.e. it wasn’t the parents but the children (who went to

In 1, the topic in the given context is idea of “the children”, and the
comment or what is asserted is that they “went to school”. Following
Lambrecht, the stress or prosodic focus on “SCHOOL” counts for the whole
phrase “went to school” as being asserted.

In the context of 2 there is a general discourse topic of “everybody”
which is presupposed. Speaker B has two jobs to do, the first of which is
to establish his topic, that is to state what he wants the hearer to take
as common ground (presupposition) which is “the children” (a subset of
“everybody”). This he does by means of stress and the full noun phrase (as
opposed to using a pronoun like “they” which would have been ambiguous).
And then, in the same clause, he goes on to his second job, that of making
a comment about that topic and asserting that they “went to SCHOOL”. The
second sentence, “GRANDMOTHER stayed HOME” is much the same, but
“GRANDMOTHER” functions to “change the topic” (from “the children”).
Levinsohn calls this a “point of departure”.

Examples 3 and 4 involve not topic-comment structures but what Levinsohn
and Lambrecht call “focus-presupposition”. In these examples, the
presupposition or common ground is an entire proposition rather than
simply the idea of an “entity” or “idea” like Children or Grandmother.
(Technically this is a presupposed proposition, and not a topic, but
remember that topic involves a kind of presupposition, so the two are very
similar.) In 3, the assertion involves “filling in the blank” and
answering “who” it was who went to school, and in 4 the assertion actually
involves the speaker correcting a false element of a presupposed
proposition, i.e. asserting that it was “the Children” and not “the
Parents” who went to school.

Now hopefully you will have noticed that in all but example 4B2 one and
the same sentence has been used, i.e. “The children went to school”. That
is to say that, grammatically and semantically (meaning-wise) each
sentence says the say thing. But what differs in each is the intonation,
and the kinds of contexts each occurs in. The claim is that the discourse
“packaging” differs in each case, and thus, in terms of presupposition
and assertion, each version is saying something different, i.e. they do
not all mean the same thing. You’ll notice though that 4B2 does the same
job more or less as 4B1, but that 4B2 involves a different kind of
sentence, what is called a “it-cleft”. In English such it-clefts are
usually found in discourse contexts where someone is correcting a false
belief, so it has a special function.

Now getting back to Greek, it is clear that one of the jobs of word order
is to indicate information structure. So when we read a commentary or
grammar that says such and such a word is “emphasized” we can stop and ask
ourselves, does it have to do with point of departure/topic and
presupposition, or does it have to do with focus, i.e. the new asserted
idea, or an asserted correction (to a presupposed false belief). (There
are other information structure categories besides these.) Most
importantly, it is helpful to keep in mind that, just as sentence stress
in English can be used to indicate a new topic as well as the focus of
assertion, so in Greek the initial position can be used to indicate both
of these things, both topic and focus, as Levinsohn has successfully
illustrated. Levinsohn however also mentions how the especially the end of
the sentence/clause is often exploited by the rules of Greek word order to
indicate especially focus (34ff).

While there are still some bugs in the Levinsohn’s program, many of his
rules work very well, and are an improvement over the standard grammars. I
think that taking the time to work through his book (or “Coursebook” as
the subtitle says) is really helpful for both relatively new as well as
advanced students (not to mention the experts). Besides information
structure proper, Levinsohn touches on a lot of other items pertaining to
the discourse structure of NT Greek, including the use of conjunctions
like KAI, DE, GAR, and conversation/quote, and how to know when one
section or “paragraph” ends and the next begins, among other issues.

I hope this apologetic helps.

Nick Bailey

B-Greek home page: http://metalab.unc.edu/bgreek
You are currently subscribed to b-greek as: [jwrobie@mindspring.com]
To unsubscribe, forward this message to leave-b-greek-327Q@franklin.oit.unc.edu
To subscribe, send a message to subscribe-b-greek@franklin.oit.unc.edu

This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:37:12 EDT