|
Abstract:
People are well known to have difficulty processing early
closure sentences such as (1):
(1) She was sure that if the state prosecutor tried the
defendant accused of embezzlement would certainly crack.
The difficulty people have processing such sentences is due to
the fact that they tend to process the noun phrase ("the
defendant") following the optionally transitive verb ("tried") as
the direct object of that verb. There has been much controversy
about the nature of the breakdown evidenced in the processing of
such garden path sentences, and most of this literature
investigates this question by studying how various types of
linguistic information -- prosodic, semantic, syntactic -- affect
the processing of such sentences. This experiment investigated
whether extra-linguistic information, namely, the melodic closure
of a line of music, can have an effect on how people resolve the
temporary syntactic ambiguity of early closure sentences.
Musical passages were constructed by evenly spacing single notes
in such a way that they are perceived as two musical phrases. Early
closure sentences were synthesized using the SimpleText text
editing program for the Macintosh computer. Eighteen music/sentence
pairs were constructed. For each music/sentence pair, three items
were constructed using the DECK audio mixing program. These three
items were mixed such that the first musical phrase closed at the
point of early closure in the sentence (p1), at the point in the
sentence at which processing breakdown normally occurs (p2), and at
a point later in the sentence than that at which processing
breakdown normally occurs (p3).
(1) She was sure that if the state prosecutor tried (p1) the
defendant accused of embezzlement (p2) would certainly (p3)
crack.
Participants in the experiment were fifteen students at San
Francisco State University. The MacLaboratory program was used to
present the stimuli to participants through headphones. Each
participant was presented with one version of each of the eighteen
music/sentence pairs -- six with musical closure at (p1), six with
musical closure at (p2), and six with musical closure at (p3) --
together with 15 clearly grammatical and 15 clearly ungrammatical
items all of which were mixed with music like the experimental
items. Participants were instructed to rate how grammatical they
thought each sentence sounded on a scale from 1 (completely
ungrammatical) to 7 (completely grammatical).
Participants rated the experimental items for which the first
musical phrase closed at the point of early closure in the sentence
(p1) as significantly more grammatical than experimental items for
which the first musical phrase closed at the point in the sentence
at which processing breakdown normally occurs (p2), suggesting that
the musical closure had an effect on the resolution of the
temporary syntactic ambiguity in early closure sentences. The
implications of these results for theories of sentence processing
will be discussed.
|
