|
Abstract:
Recent psycholinguistic studies have paid increasing
attention to the role of prosody in syntactic disambiguation, but
little work has been done on prosodic disambiguation of semantic
structure. This paper presents the results of a production
experiment and a comprehension experiment on the role of prosody in
disambiguating scope relations in Greek sentences (see 1, 2 below).
The linguistics literature treats scope relations as the direct
result of c-command: A scopes over B if A c-commands B, providing
meaning A of the examples. Under this view, the derivation of
inverse scope (B scoping over A), meaning B of the examples, is
problematic because c-command relations between A and B need to be
reversed for the right interpretation.
Syntactic-semantic theory approaches this problem with syntactic
movement mechanisms which build extra structure which is both
uneconomical and stipulative, but nevertheless fail to derive all
and only the correct interpretations: in some cases they generate
more interpretations than a given sentence admits to and in some
cases fewer.
The present studies examine whether prosodiceffects can better
explain scope interpretations. In a production experiment, five
Athenian Greek speakers, naïve to the purposes of the
experiment, each read aloud seven sentences embedded in
disambiguating contexts. The sentences contained negation and
another operator: two sentences (example (1)) contained "because",
and five sentences (example (2)) contained a quantified NP (eg "90
barrels"). The 35 resulting unambiguous tokens were then subjected
to extensive prosodic analyses to determine their prosodic
structure. The results show significant differences in the prosodic
structure of the two interpretations: inverse scope, as in meaning
B below, resulted if there was a prosodic boundary after negation,
while linear scope, as in meaning A, resulted if there was no
boundary after negation. In the comprehension study, the 35 target
sentences were played to 40 Greek listeners who were given a
forced-choice paraphrase selection task. The results showed
significantly higher than chance levels of correct categorization
of the sentences.
These results suggest that there is a strong correlation between
prosodic cues and scope disambiguation in Greek ---and intuitively
one would expect the same kind of results to obtain in English, as
well. More specifically, we found that a prosodic boundary
demarcated the scope of negation: if there was a boundary after
negation, then the operators had inverse scope as in meaning B, and
if there wasn't a boundary, then the operators had linear scope, as
in meaning A. Both speakers and listeners seem to unconsciously
make use of these prosodic cues whenever they are interpreting
sentences. I will offer thoughts on possible explanations for this
phenomenon during my presentation.
Examples
(1) den gelai giati zilevi
not laughs because is-jealous = "He doesn't laugh because he's
jealous"
Meaning A: not > because: "He laughs but not because he's
jealous; for some other reason"
Meaning B: because > not: "His jealousy prevents him from
laughing"
(2) den gemizun enenida varelia
not fill-3pl. ninety barrels = "They don't fill ninety barrels"
Meaning A: not > 90 barrels: "The number of filled barrels is
not 90"
Meaning B: 90 barrels > not: "As for 90 barrels, they don't
fill them"
|
