sentences

Carol Genetti cgenetti at HUMANITAS.UCSB.EDU
Mon Jun 29 16:51:17 UTC 1998


Dear funknetters,

        Pardon me for coming in on this discussion late -- with a six
month old baby and two students about to leave for their first field trip
to Nepal, I've been deleting list mail without reading it.  BUT, I'm very
interested in the notion of sentence and would like to put in my two cents
worth.

        I work on Indo-Aryan and Tibeto-Burman languages of the Himalayas.
These languages are all verb-final, and have the standard features that
correlate with verb-final order, especially "clause chaining" or "converb
constructions" (whether or not these are really separate is an interesting
question).  In the languages I have looked at extensively, Dolakha Newar
and Nepali, sentence boundaries are quite clear, marked by the presence of
a finite verb, after which may be a number of discourse particles and/or
postposed elements.  The structure of these sentences may be quite
complex, entailing multiple levels of embedding, as well as chaining
structures. It is clear that the sentences are unified wholes and that
they are significant syntactic units that speakers attend to and
manipulate as they produce spoken discourse.

        I've written two papers (both in press) which examine different issues
of sentencehood.  (Both papers examine narrative -- conversational data is
very interesting, and units are more likely to be left incomplete -- I
still think that the notion of sentence is relevant there as well.)Both
papers give a lot of syntactic argumentation for what a sentence is,
particularly with reference to embedded quotation. The two papers haVE
different emphases, goals, and languages. In one, co-authored with Keith
Slater, we looked at sentences, clause boundaries and intonation.  We
propose that there are "prosodic sentences", consisting of a series of
prosodic units with non-final intonation, and ending in a unit with final
intonation (analogous to clause-chaining structures).  Prosodic sentences
may be internally complex and involve embedding.  Prosodic sentences and
syntactic sentences generally co-terminate, but there a number of other
interesting patterns as well. We also use the term "narrative sentence"
which seem to be the units speakers most clearly delineate, but which may
not have final marking at either the prosodic or syntactic levels. This
paper contains a complete narrative (a folk rendition of the beginning of
the Mahabharata) intonationally transcribed, glossed, and extensively
annotated.  If anyone wants a copy, just let me know!

        A second paper which concerns the sentence I co-authored with
Laura Crain.  It looks at issues of preferred argument structure in
Nepali, and demonstartates that the amount and type of nominal reference
is based not on the clause in this language, but on the sentence.  We
found that speakers have a preference to make one overt mention of each
referent one time in a sentence, regardless of the number of times the
referent occurs as a verbal argument in the sentence (and regardless of
the discourse prominence of the referent). Thus sentences are key units
that speakers are aware of and manipulate. This paper backs up the central
claims of PAS theory, namely that grammar and discourse patterns are
complementary, and shows that the typological facts of Nepali favor the
discourse patterns found. This paper is to appear in Du Bois et al. Again,
if anyone is interested in obtaining a copy, just let me know.

        So, regarding the relevance of the sentence, I think language
typology is a significant factor.  Languages differ considerably in their
syntactic patterns, and the relevance of any syntactic unit, in particular
the sentence, will vary with the typology.


        -- Carol Genetti



More information about the Funknet mailing list