Prosodic Effects on Relative Clause Attachment

1
Prosodic Effects on Relative Clause Attachment

• Work in progress, hopefully almost done.

2
Schafer et al., 1996

• The sun was shining on
• the propeller of the plane
• that the mechanic was
  carefully
• 
  examining.

3
Schafer et al., 1996

• The sun was shining on
• H
• the propeller of the plane i.p.
• 
  H
• the propeller of the plane i.p.
• that the mechanic
  was carefully
• 
  examining.

4
Maynell, 2001

• Test LH pitch accents in combination with
  presence or absence of IP boundary.
• Test wh-relatives and that-relatives.

5
Example Stimuli

The investigator found the uncle of the
businessman who was wanted by police

The investigator found the uncle of
the businessman who was wanted by police
The
investigator found the uncle of the businessman
that was wanted by police

The investigator found the uncle of the
businessman that was wanted by police
The
collector admired the leg of the table which
was beautifully carved.
The collector
admired the leg of the table which was
beautifully carved.
The collector admired the leg of
the table which was beautifully carved.

The collector admired the leg of the table
which was beautifully carved.
6
N1 accented, IP boundary before Relative Clause
7
N2 accented, IP boundary before Relative Clause
8
N1 accented, no IP boundary before Relative Clause
9
N2 accented, no IP boundary before Relative Clause
10
Experimental Design

• Phrase boundary conditions (between subj)
• IP, 500ms silence
• IP, 0ms silence
• no boundary
• Accent conditions (within subjects)
• N1 accented
• N2 accented

11
Task

• Participants listened to sentences played on a
  tape recorder
• Each sentence presented twice
• Answered the corresponding question on a paper
  answer sheet by circling one of two possible
  choices.
• Were not allowed to look at question until the
  next sentence began playing

12
Answer Sheet

• 1. What color was it painted?
• A. Blue. B.Green _____________________________
  ___________________
• 2. What indicated that there was something
  wrong?
• A. A light was on. B. The monitor was
  glowing.
• __________________________________________________
  _
• 3. What was five years old?
• A. The alarm. B. The house.

13
Results wh Relatives
14
Results that-Relatives
15
Summary of Results

• Main effect of accent (though not for wh-RCs)
• Huge main effect of IP boundary
• No effect of 0 ms vs. 500 ms silence
• Effect of pronoun type wh-RCs show higher
  attachment than that-RCs

16
Written Experiment

• Purpose
• To determine whether commas in written language
  will have the same effect as IP boundaries in
  spoken language on the resolution of the relative
  clause ambiguity.

17
Design
18
Task

• Participants read test and filler sentences in
  order
• Instructed to write first answer that came to
  mind
• Told not to change answers

19
Example responses

• The uncle (N1)
• The businessman (N2)
• The uncle of the businessman (N1)
• Either the uncle or the businessman (ambig)
• The investigator (wrong)

20
Results
21
Results Items Analysis

• Main effect of punctuation
• F(1, 15) 25.053, plt.001
• Main effect of relative clause type (that/wh)
• F(1,15) 11.087, plt0.01
• No interactions

22
Remaining Work

• Run analysis by subjects
• Clean up statistics from auditory experiment
• Revise paper
• Move on

23
THANK YOU!

• The End.