Post on 28-Dec-2015
Intonational prominence:Intonational prominence:Production in dialogue, Production in dialogue, comprehension during comprehension during
visual search visual search
Shari R. Speer (with much input from Shari R. Speer (with much input from Kiwako Ito, Co-I)Kiwako Ito, Co-I)
Ohio State University LinguisticsOhio State University Linguistics
Intonation & Pragmatic StatusIntonation & Pragmatic Status
English Pitch AccentEnglish Pitch Accent
Distinctive tonal movement associated with Distinctive tonal movement associated with stressed syllable. stressed syllable.
perceivable prosodic prominenceperceivable prosodic prominence
Local targets L and H (Pierrehumbert 1980)Local targets L and H (Pierrehumbert 1980)
H*, (H+!H*), L*, L+H*, L*+HH*, (H+!H*), L*, L+H*, L*+H
(a/c current ToBI guideline)(a/c current ToBI guideline)
Pitch Accent Types & PragmaticsPitch Accent Types & Pragmatics
Pierrehumbert & Hirschburg, 1990Pierrehumbert & Hirschburg, 1990Compositional tune meaningCompositional tune meaningPA typePA type meaningmeaningH*H* New items introduced to a New items introduced to a
discourse, added to mutual belief space. discourse, added to mutual belief space.
L*L* Already part of mutual beliefs.Already part of mutual beliefs.Salience-without-predicationSalience-without-predication
L*+HL*+H Uncertainty, lack of speaker commitmentUncertainty, lack of speaker commitment
L+H*L+H* Contrast with some alternative related item Contrast with some alternative related item
Pierrehumbert & Hirschberg, 1990Pierrehumbert & Hirschberg, 1990Compositional tune meaningCompositional tune meaningPA typesPA types meaningmeaningH*H* New items introduced to a New items introduced to a
discourse, added to mutual belief space.discourse, added to mutual belief space.
L*L* Already part of mutual beliefs.Already part of mutual beliefs.Salience-without-predicationSalience-without-predication
L*+HL*+H Uncertainty, lack of speaker commitmentUncertainty, lack of speaker commitment
L+H*L+H* Contrast with some alternative related itemContrast with some alternative related item
Pitch Accent Types & PragmaticsPitch Accent Types & Pragmatics
Categorical distinction Categorical distinction between H* and L+H*between H* and L+H*
F0 minima preceding the accentual peak should be stable for F0 minima preceding the accentual peak should be stable for L+H*, but not for H*L+H*, but not for H*
PerceptionPerception
A clear categorical boundary should be observed in both A clear categorical boundary should be observed in both identification and discrimination task along with H* height identification and discrimination task along with H* height continuum.continuum.
Tonal shapeTonal shape
H*H* L+H*L+H*
flat or subtle riseflat or subtle rise clear riseclear rise
Categorical distinction Categorical distinction between H* and L+H*between H* and L+H*
F0 minima between two successive H* are consistently F0 minima between two successive H* are consistently aligned to the syllable onset (Ladd & Schepman, 2003). aligned to the syllable onset (Ladd & Schepman, 2003).
Listeners use the F0 minima alignment cue to distinguish Listeners use the F0 minima alignment cue to distinguish ambiguous names, e.g. Norma Nelson vs. Norman Elson ambiguous names, e.g. Norma Nelson vs. Norman Elson (Ladd & Schepman, 2003).(Ladd & Schepman, 2003).
Although context identification tasks show categorical Although context identification tasks show categorical distinction between non-emphatic vs. emphatic contexts, distinction between non-emphatic vs. emphatic contexts, listeners perform poorly in discrimination tasks along the listeners perform poorly in discrimination tasks along the accentual peak continuum. (Ladd & Morton, 1997)accentual peak continuum. (Ladd & Morton, 1997)
Categorical distinction Categorical distinction between H* and L+H*between H* and L+H*
Testing tonal shapeTesting tonal shape(Arvaniti & Garding, to appear)(Arvaniti & Garding, to appear) The height of peak and its alignment vary The height of peak and its alignment vary
according to the degree of emphasis. according to the degree of emphasis. Dialectal differences: Dialectal differences:
Contrast between H* & L+H* preserves in Contrast between H* & L+H* preserves in Southern Californian but not in Minnesotan.Southern Californian but not in Minnesotan.
H* <--> L+H*H* <--> L+H* ““gradient variationgradient variation between two ends of a between two ends of a continuum” continuum” (Ladd & Schepman, 2003)(Ladd & Schepman, 2003)
Analyzing Unscripted SpeechAnalyzing Unscripted Speech
Task: Give instructions for tree decoration Task: Give instructions for tree decoration Monitor display with a picture of ornament Monitor display with a picture of ornament
with its tag (e.g. blue ball) and a picture of with its tag (e.g. blue ball) and a picture of a tree, indicating the location of the a tree, indicating the location of the ornament.ornament.
Newness, Givenness & Contrastiveness Newness, Givenness & Contrastiveness manipulated by the order of mention. manipulated by the order of mention.
*
*
*
*
*
16 17
18 19 20 21 22
23 24 25 26 27
28 29 30 31 32
*
* *
* *
* * *
*orange orange candycandy
*
*
*
*
*
17
18 19 20 21 22
23 24 25 26 27
28 29 30 31 32
*
* *
* *
* * *
* *blue blue candycandy
*
*
*
*
*
green green candycandy
18 19 20 21 22
23 24 25 26 27
28 29 30 31 32
*
* *
* *
* * *
* * *
*
*
*
*
*
16 17
18 19 20 21 22
23 24 25 26 27
28 29 30 31 32
*
* *
* *
* * *
*orange orange candycandy
*
*
*
*
*
17
18 19 20 21 22
23 24 25 26 27
28 29 30 31 32
*
* *
* *
* * *
* *blue blue candycandy
*
*
*
*
*
green green candycandy
18 19 20 21 22
23 24 25 26 27
28 29 30 31 32
*
* *
* *
* * *
* * *
Example pitch tracesExample pitch traces
Time (s)0 1.19152
75
125
Time (s)0 1.13119
75
125
Accentuation ProportionAccentuation Proportion
NEW/NEW/
GIVENGIVEN# OF # OF
TRIALSTRIALS ADJECTIVEADJECTIVE NOUNNOUN
NewNewNewNew 3131 .94.94 .84.84
NewGivenNewGiven 3232 .84.84 .50.50
GivenNewGivenNew 3232 .91.91 .81.81
GivenGivenGivenGiven 9696 .89.89 .67.67
Accentuation ProportionAccentuation Proportion
NEW/NEW/
GIVENGIVEN# OF # OF
TRIALSTRIALS ADJECTIVEADJECTIVE NOUNNOUN
NewNewNewNew 3131 .94.94 .84.84
NewGivenNewGiven 3232 .84.84 .50.50
GivenNewGivenNew 3232 .91.91 .81.81
GivenGivenGivenGiven 9696 .89.89 .67.67
L+H* distributionL+H* distribution
NEW/NEW/
GIVENGIVEN
PRAGMATICPRAGMATIC
STATUSSTATUS
ADJECTVEADJECTVE NOUNNOUN
# OF # OF TRIALSTRIALS L+H*L+H* # OF # OF
TRIALSTRIALS L+H*L+H*
NewNewContrastiveContrastive 2626 .46.46 3131 .19.19
Non-Non-contrastivecontrastive 3737 .03.03 3333 .06.06
GivenGivenContrastiveContrastive 3737 .51.51 3333 .18.18
Non-Non-contrastivecontrastive 9191 .00.00 9595 .00.00
Is intonation informative?Is intonation informative?Effect of Intonation on visual searchEffect of Intonation on visual searchListeners use pitch accent to single out the object referred Listeners use pitch accent to single out the object referred
to (Dahan et al. 2002)to (Dahan et al. 2002)
1. “Put the candle/candy below the triangle.”1. “Put the candle/candy below the triangle.”
2.2. “NOW, put the “NOW, put the CANDLECANDLE above the square.” above the square.” “ “NOW, put the NOW, put the candlecandle ABOVE THE SQUARE.” ABOVE THE SQUARE.”
fixation to fixation to candycandyfixation to fixation to candlecandle
candle candle “ “CANDLECANDLE” ” candle candle “ “candycandy” ”
Research QuestionsResearch Questions
How does intonational prominence (L+H*) on a How does intonational prominence (L+H*) on a modifier adjectivemodifier adjective affect visual search? affect visual search?
L+H*L+H*: Evokes a contrast set that contains alternative related : Evokes a contrast set that contains alternative related item(s)item(s)
(Pierrehumbert & Hirschberg, 1990)(Pierrehumbert & Hirschberg, 1990)
BLUEBLUE ball ball
Research QuestionResearch Question
How does intonational prominence (L+H*) on a How does intonational prominence (L+H*) on a modifier adjectivemodifier adjective affect visual search? affect visual search?
L+H*L+H*: Evokes a contrast set that contains alternative related : Evokes a contrast set that contains alternative related item(s)item(s) (Pierrehumbert & Hirschberg, 1990)(Pierrehumbert & Hirschberg, 1990)
BLUEBLUE ballballgreen ballgreen ball
Why track the eyes?Why track the eyes? Eyemovements can be traced:Eyemovements can be traced:
during natural conversationduring natural conversation for both speaker & listenerfor both speaker & listener to investigate the online language processing to investigate the online language processing
(Tanenhaus and Trueswell,2005)(Tanenhaus and Trueswell,2005)
Accentual prominence evokes reference to an alternative Accentual prominence evokes reference to an alternative discourse entity (Dahan et al., 2002).discourse entity (Dahan et al., 2002).
Listeners fixate target objects 200ms from the onset of Listeners fixate target objects 200ms from the onset of auditory stimuli during simple visual search (Allopenna, auditory stimuli during simple visual search (Allopenna, et al. 1998). As much as 150ms of this time is used to et al. 1998). As much as 150ms of this time is used to program and execute a ballistic saccade (Matin, Shao, & program and execute a ballistic saccade (Matin, Shao, & Boff, 1993).Boff, 1993).
Experiment 1:ProcedureExperiment 1:Procedure Subjects followed recorded audio instruction to decorate Subjects followed recorded audio instruction to decorate
Christmas trees. Christmas trees.
(ASL e5000) at 60Hz.(ASL e5000) at 60Hz.
Experiment 1: Visual StimuliExperiment 1: Visual Stimuli
Ornaments displayed on a grid with 11 cells (8 Ornaments displayed on a grid with 11 cells (8 target + 3 filler)target + 3 filler)
Experiment 1: Sound StimuliExperiment 1: Sound Stimuli
Accentual patterns of stimuli were selected Accentual patterns of stimuli were selected based on the analysis of spontaneous speech based on the analysis of spontaneous speech (Ito, Speer & Beckman, 2003).A trained female (Ito, Speer & Beckman, 2003).A trained female phonetician produced stimuli with the intended phonetician produced stimuli with the intended Pitch accent patterns.Pitch accent patterns.
Recorded 44.1KHz, 16bitsRecorded 44.1KHz, 16bits F0 extracted with 10ms windowF0 extracted with 10ms window All stimuli were ToBI transcribed by two native All stimuli were ToBI transcribed by two native
speakers of American English. speakers of American English.
Experiment 1: StimuliExperiment 1: Stimuli
Adjective NounAdjective Noun Accent PatternAccent Pattern
New New NewNew H* H* !H*!H*
New New GivenGiven H*H* !H*!H*
Given Given GivenGiven H*H* !H*!H*
GivenGiven NewNew H*H* H*H*
NewCNewC GivenGiven L+H*L+H* no accent no accent
GivenGiven NewCNewC H*H* L+H*L+H*
Experiment 1: StimuliExperiment 1: Stimuli
Felicitous PatternsFelicitous Patternsgreen onion green onion ORANGEORANGE onion onion
L+H* L+H* no accentno accent brown ball brown ball brown brown ANGELANGEL
H* H* L+H* L+H* Infelicitous PatternsInfelicitous Patterns
gray stockinggray stocking brown brown STOCKINGSTOCKING H* H* L+H* L+H*
orange candy orange candy ORANGEORANGE onion onion L+H* L+H* no accentno accent
Experiment 1: StimuliExperiment 1: Stimuli
Time (s)0 1.44748100
400
theorangeonion0L+H*0noL–L%1 1 4HTime (s)0 1.44748
Time (s)0 2.04925100
400
nexthangthebrownangelH*H–00H*0L+H*L–L%3 1 1 4H HLTime (s)0 2.04925
All stimuli were ToBI-transcribed by two native All stimuli were ToBI-transcribed by two native speakers of American English.speakers of American English.
Ave. 300 Hz Ave. 299 Hz
Experiment 1: ResultsExperiment 1: Results
[H* !H*] NewNew vs. NewGiven
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations to target
NewNew
NewGiven
Experiment 1: ResultsExperiment 1: Results
[H* !H*]: NewNew vs. GivenGiven
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations to target
NewNew
GivenGiven
Experiment 1: Given/New Experiment 1: Given/New
Objects named by given nouns were fixated Objects named by given nouns were fixated sooner and more often than objects named by sooner and more often than objects named by new nouns. new nouns.
No additional advantage came from having both No additional advantage came from having both a given adjective and a given noun (compared to a given adjective and a given noun (compared to difference for given noun alone).difference for given noun alone).
Visual search facilitated by previous mention of Visual search facilitated by previous mention of object nouns, and previous successful search object nouns, and previous successful search for the referent in object-organized array.for the referent in object-organized array.
Experiment 1: ResultsExperiment 1: Resultsgreen onion ORANGE onion [L+H* noacc]
orange ONION [H* L+H*]
Contrast on Color: [L+H* no accent] vs. [H* L+H*]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations to target
L+H* no accent:Felicitous
H* L+H*:Infelicitous
QuickTime™ and aH.263 decompressor
are needed to see this picture.
Experiment 1: ResultsExperiment 1: Results
Contrast on Object: [H* L+H*] vs. [L+H* no accent]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations to target
H* L+H*:Felicitous
L+H* no accent:Infelicitous
brown drum brown ANGEL BROWN angel
Experiment 1: Felicitous vs. Experiment 1: Felicitous vs. InfelicitousInfelicitous
Felicitous L+H* on the color term in contrastive Felicitous L+H* on the color term in contrastive environments (blue egg, SILver egg) facilitated environments (blue egg, SILver egg) facilitated visual search compared to infelicitous L+H* on visual search compared to infelicitous L+H* on non-contrastive noun (grey candy, gold CANdy). non-contrastive noun (grey candy, gold CANdy).
Felicitous L+H* on the object noun in contrastive Felicitous L+H* on the object noun in contrastive environments (blue egg, blue CANdy) showed environments (blue egg, blue CANdy) showed no advantage over infelictous L+H* on the no advantage over infelictous L+H* on the adjective (blue egg, BLUE candy).adjective (blue egg, BLUE candy).
Listeners ‘tune’ their sensitivity to contrastive Listeners ‘tune’ their sensitivity to contrastive accent on the basis of the visual task?accent on the basis of the visual task?
Experiment 2: QuestionsExperiment 2: Questions
Does Does L+H*L+H* on the contrastive adjective lead to on the contrastive adjective lead to anticipatory eye-movement compared to anticipatory eye-movement compared to H*H*? ? e.g. brown drum --> e.g. brown drum --> REDRED//redred drum drum
Does infelicitous Does infelicitous L+H*L+H* on the color adjective lead on the color adjective lead to an expectation strong enough to lead a to an expectation strong enough to lead a garden-path in eye-movement?garden-path in eye-movement?
e.g. red onion --> e.g. red onion --> GREENGREEN drum drum
Experiment 2Experiment 2
Procedure Same as Exp 1Procedure Same as Exp 1 Subjects followed audio instruction to decorate Subjects followed audio instruction to decorate
Christmas trees. Christmas trees. Ornaments displayed on a grid with 8 target + 3 Ornaments displayed on a grid with 8 target + 3
filler cellsfiller cells Eye-movements monitoring at 60Hz.Eye-movements monitoring at 60Hz.
StimuliStimuli Prepared in the same way as Exp1Prepared in the same way as Exp1 Same speaker and ToBI transcribersSame speaker and ToBI transcribers
Experiment 2: ResultsExperiment 2: Results
Contrast on Color: [L+H* no accent] vs. [H* !H*]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations to target
L+H* no accent:Felicitous
H* !H*:Infelicitous
brown drum RED drum [L+H* no acc] red drum [H* !H*]
Experiment 2: ResultsExperiment 2: Results
Previously Mentioned Adjective & Noun: [H* !H*] vs. [L+H* no accent]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations to target
H* !H*:Felicitous
L+H* no accent:Infelicitous
red onion green drum GREEN drum
Experiment 2: ResultsExperiment 2: Results
Infelicitous [L+H* no accent]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations
Looks to targetLooks to previous target
red onion GREEN drum [L+H* no acc]
onion
drum
The effect of infelicitous L+H*The effect of infelicitous L+H*
QuickTime™ and aDV/DVCPRO - NTSC decompressor
are needed to see this picture.
Experiment 2: ResultsExperiment 2: Results
Infelicitous [L+H* no accent]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations
Looks to targetLooks to previous target
red onion GREEN drum [L+H* no acc]
onion
drum
Experiment 2: ResultsExperiment 2: Results
Felicitous [H* !H*]
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 300 600 900 1200 1500 1800 2100 2400 2700 3000
Milliseconds from noun onset
Proportion of fixations
Looks to target
Looks to previoustarget
red onion green drum
onion
drum
Experiment 2: SummaryExperiment 2: Summary For cases with repetition of the immediately preceding For cases with repetition of the immediately preceding
noun, L+H* showed a processing advantage as compared noun, L+H* showed a processing advantage as compared to H* on the adjective (brown drum --> RED/red drum)to H* on the adjective (brown drum --> RED/red drum)
For cases without repetition of the noun, infelicitous L+H* For cases without repetition of the noun, infelicitous L+H* led to led to incorrect anticipatory fixationsincorrect anticipatory fixations to the most-recently to the most-recently mentioned ornament type.mentioned ornament type.
significant delay in fixations to the real target significant delay in fixations to the real target
NO incorrect anticipatory fixationsNO incorrect anticipatory fixations were were observed when observed when non-repeated non-repeated targets were targets were felicitously presented with [H* !H*] felicitously presented with [H* !H*]
Experiment 2: SummaryExperiment 2: Summary
DiscussionDiscussion Dahan et al results showed that L+H* accent led to the Dahan et al results showed that L+H* accent led to the
expectation of contrast. Contrastive accent led listeners expectation of contrast. Contrastive accent led listeners to fixate the object type that had *not* just been to fixate the object type that had *not* just been mentioned. mentioned.
Results here show contrast effects from L+H* accent on Results here show contrast effects from L+H* accent on the modifier that preceded the noun. L+H* led listeners to the modifier that preceded the noun. L+H* led listeners to restrict the set of expected referents to those of the restrict the set of expected referents to those of the object type that *had* been most recently mentioned. object type that *had* been most recently mentioned.
This evidence suggests that L+H* evokes generation of a This evidence suggests that L+H* evokes generation of a contrast set in the discourse representation of the contrast set in the discourse representation of the listener. The set is based on the accented word listener. The set is based on the accented word interpreted within the structure of the utterance. Here, interpreted within the structure of the utterance. Here, contrastive accent on the adjective restricted contrastive accent on the adjective restricted expectations about the coming head noun to the set of expectations about the coming head noun to the set of objects specified by the immediately preceding noun.objects specified by the immediately preceding noun.
Conclusions so farConclusions so far
Naïve speakers predictably used L+H* accents to convey Naïve speakers predictably used L+H* accents to convey contrast when directing a visual search during spontaneous contrast when directing a visual search during spontaneous dialogue. dialogue.
Eyemovement latencies and fixation probabilities showed Eyemovement latencies and fixation probabilities showed listeners’ immediate use of pitch accent information during listeners’ immediate use of pitch accent information during discourse comprehension.discourse comprehension.
Listeners were sensitive to the relevance of intonational cues Listeners were sensitive to the relevance of intonational cues for determining the intended referent during visual search.for determining the intended referent during visual search.
L+H* accents on adjectives evoked an immediate interpretation L+H* accents on adjectives evoked an immediate interpretation of contrast, which led to anticipatory fixations to the most-of contrast, which led to anticipatory fixations to the most-recently mentioned object type.recently mentioned object type.
These effects were not due to simple salience of the L+H* These effects were not due to simple salience of the L+H* accented item, nor to word duration effects. accented item, nor to word duration effects.
Additional studiesAdditional studies
““Christmas Tree” production and Christmas Tree” production and Eyemovement studies with Japanese Eyemovement studies with Japanese speakersspeakers
Further investigation of discourse structure Further investigation of discourse structure and the distribution of pitch accent type in and the distribution of pitch accent type in EnglishEnglish
Additional studies examining other aspects Additional studies examining other aspects of intonational effects on discourse of intonational effects on discourse structure, e.g. structure, e.g. Deaccentuation (reduction, deletion, pronouns) in Deaccentuation (reduction, deletion, pronouns) in
production and comprehensionproduction and comprehension
THANK YOU!THANK YOU!Acknowledgements:Acknowledgements:
Ping BaiPing Bai
Allison BlodgettAllison Blodgett
Laurie MaynellLaurie Maynell
Mary BeckmanMary Beckman
NSF BCS-018464NSF BCS-018464
NIDCD R01-DC007090NIDCD R01-DC007090
Experiment 1: Pre-fixation data, Experiment 1: Pre-fixation data, Felicitous/Infelicitous Felicitous/Infelicitous
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
-600 -300 0 300 600 900 1200 1500
Time from noun onset (ms)
Fixation proportion to the target object cell
Felicitous Adj Contrast BLUE drum
Infelicitous Adj Contrast blue DRUM
Felicitous Noun Contrast blue DRUM
Infelicitous Noun Contrast BLUE drum
blue
BLUE
DRUM
drum
Experiment 2: Pre-fixation data, Experiment 2: Pre-fixation data, Felicitous/Infelicitous Felicitous/Infelicitous
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
-600 -300 0 300 600 900 1200 1500
Time from noun onset (ms)
Fixation proportion to the target object cell
Felicitous Adj Contrast GREEN drum
Infelicitous Adj Contrast green drum
Felicitous Both Already green drum
Infelicitous Both Already GREEN drum
GREEN drum
green drum
Experiment 1: Additional Experiment 1: Additional findingsfindings
Overall, shorter latencies for trials with immediately repeated object Overall, shorter latencies for trials with immediately repeated object nouns nouns
Overall, shorter latencies for L+H* vs. H* trialsOverall, shorter latencies for L+H* vs. H* trials Comparisons for ‘mismatch’ cases are between felicitous and infelicitous Comparisons for ‘mismatch’ cases are between felicitous and infelicitous
L+H* in different utterance positionsL+H* in different utterance positions Additional comparison needed:Additional comparison needed:
For trials with repeated nounsFor trials with repeated nouns
Felicitous [L+H* no accent] vs. Infelicitous [H*!H*]Felicitous [L+H* no accent] vs. Infelicitous [H*!H*]
For trials with non-repeated nounsFor trials with non-repeated nounsFelicitous [H*!H*] vs. Infelicitous [L+H* no accent]Felicitous [H*!H*] vs. Infelicitous [L+H* no accent]
For trials with L+H* in the same word positionFor trials with L+H* in the same word positionFelicitous L+H* no accent vs. Infelicitous L+H* noaccentFelicitous L+H* no accent vs. Infelicitous L+H* noaccent