What does the motor theory state about speech perception?
Speech is perceived by the processes that are also involved in its production *motor commands
What is a weak version of motor theory?
Speech production offers important clues about speech perception that can be used by listeners *More appropriate*
What is a strong version of motor theory?
Speech production forms the BASIS for speech perception
What are some drawbacks to the motor theory?
1) We don't know how a person makes the link between acoustic speech signals and motor commands in a speaker's brain 2) Why do motor commands have to be the mediator between acoustic signal and phonemes? It should be able to go directly to phonemes.
What is an EMG?
Allows you to look at the neural signals to the muscles of articulation: records electrical impulses triggering muscle contractin
For an EMG, electrodes placed ____ provide most accurate readings.
Directly into the muscles
What did EMG studies show regarding motor theory?
EMG traces were as variable as the acoustic signal-NO support for strict motor theory
What is some evidence related to aphasia that goes against motor theory?
1) Pts with Broca's Aphasia have motor problems in producing speech but do not usually have as difficult of a problem with perception 2) Pts with Wernicke's Aphasia do not have motor problems but have more issues with speech perception
However, the speech perception abilities of Wernicke's aphasics have ___ been shown to be statistically inferior to those of Broca's.
What is some evidence to support the link between perception and production (motor theory)?
1) Foreigners moving their lips to understand unfamiliar languages, "speech reading" 2) McGurk Effect
What is the McGurk Effect?
When listeners combine visual information on speech production with auditory information to perceive a place of articulation that is in BETWEEN the two delivered by seeing and listening
What is an example of the McGurk effect?
-/ba/ is dubbed into lip movements for /ga/ but adults reported hearing /da/ -"Talking" animals on TV
In a weakened version of motor theory, speech perception is not ____.
driven by motor commands
What does the revised motor theory take into account?
Modules and Duplex Perception
What is the idea behind "modules"?
The brain has developed specialized areas for treating certain perceptual info-act INDEPENDENTLY of other brain processing
What is duplex perception?
Simultaneously processing the same acoustic information in both a speech and non-speech mode
What is analysis-by-synthesis theory?
*Again: perception based on production -Listeners decode acoustic signal by internally generating a matching signal. The one that provides the best match is the one perceived by the listener.
What is the most important difference bw analysis by synthesis and motor theory?
Analysis by Synthesis does not mandate that it is specifically motor commands that mediate perception
What are templates?
The concept of internalized recognition patterns "stencils"
The concept of templates or internalized recognition patterns is borrowed from the study of ___.
What is a drawback to analysis by synthesis?
1) Takes a long time to go through the matching process
What method was proposed to alleviate the lengthy matching process?
A pre-analysis: breaking the speech signal down into distinctive features
Templates can be ___ or ___.
Precise/Restrictive or General. General templates match more acoustic signals.
CH 7: What are distinctive features?
The absolute minimal contrasts between phonemes in a language
What are universal distinctive features?
-Intended to be applicable for every language -Strictly acoustic in nature and NOT based on articulation -Are binary (+/-), distinctive and minimal-represent relevant contrasts between phonemes with fewest number of features possible
How are Chomsky and Halle's features different from universal distinctive features?
Were developed for the English language--NOT universal. Mix of phonological and articulatory features
The distinctive features are optimal..they convey the ___ number of contrasts necessary.
What is the formula that accounts for the minimal number of binary contrasts necessary to convey info?
2 to the nth power n=number of binary features
Are the descriptive terms bilabial, alveolar, and velar minimal distinctive features?
What are the two binary features related to Jakobson and Halle?
The acoustic features compact and acute relate to ____ in English.
Place of articulation
How are velars classified?
+ Compact (-diffuse)
How are bilabials and alveolars classified?
How can bilabials be further classified?
-Compact -Acute +Grave
How can alveolars be further classified?
-Compact +Acute -Grave
Why don't we have to specify whether +compact consonants (velars) are acute or grave?
Because there are no contrasts between consonants that are +compact
A ____ is one way of coping with an extraordinary amount of variation found in the acoustic signal of speech.
What is a feature analysis?
Breaks down the signal in terms of its relevant acoustic properties. *Only retains contrasts relevant for speech perception
What is an example of an acoustic contrast that would be ignored after a feature analysis?
Differences due to speech being produced by different speakers
What are some examples of evidence that listeners classify speech sounds by features?
1) Listening errors in the presence of noise are highly organized (ex: place of articulation and voicing errors common) 2) Slips of the tongue-errors that can be often described as change in a single feature
What are phonetic feature detectors?
Neural centers that respond to speech and are sensitive to certain acoustic properties
Although phonetic feature detectors have not been isolated in the human brain, scientists have isolated areas in ____.
Cats and other animals
Where does the main source of evidence for phonetic feature detectors come from?
Selective adaptation procedure
What happens with the selective adaptation procedure?
A certain acoustic property detector is fatigued through repeated presentation of an auditory stimuli with that property present (ex: after repeated presentations of a voiced sound, listeners hear more stimuli as voiceless)
For selective adaptation procedure investigations, scientists obtained identification functions for stimuli that differed only in ___.
The adaptation effect is largest for stimuli that are _____.
near the original phonetic boundary
The adaptation effect is also seen in ____ functions.
Neural property detectors are ____ based.
Are phonetic feature detectors specific to speech?
Yes-when vowel portion of adapting stimulus is removed so that only formant transition remains, there is much less adaptation
What have phonetic features been investigated for?
Voicing and place of articulation contrast
The phonetic feature detector for voicing operates ____ of place of articulation.
Independently** they work at the phonetic/feature level
Phonetic feature detectors may help explain the ___ with which speech is produced.
Rapidity-also how newborns organize speech along various parameters
What is the theory of acoustic invariance?
Based upon the idea that invariance can be found directly in the acoustic signal
In the theory of acoustic invariance, spectral analysis (LPC) is applied at ____.
The burst resulting from the release of the consonant obstruction from the lips
The theory of acoustic invariance says that there is enough information in the acoustic signal at the burst and up to ___ ms afterwards to signify ___.
20ms, place of articulation
How long are formant transitions?
Under the acoustic invariance theory, there are no ___.
With acoustic invariance theory, what happens once the formant extraction analysis has been performed?
It is categorized along the acoustic features established by Jakobsen
If there is one formant peak that dominates the acoustic spectrum, it is ___ or ___.
If the formant peaks are rising in amplitude, it is considered ___ or ____.
Acute or Alveolar
If the formant peaks are declining in amplitude, it is considered ___ or ___.
Grave or Bilabial
In addition to a formant extraction analysis and acoustic features, the acoustic invariance theory makes use of ____.
The information used to determine the place of articulation in the acoustic invariance theory is ____.
What does "static" mean?
We are only looking at the acoustic signal at one particular fixed point in time
What does "dynamically" mean?
Information is gathered over a longer time period
Why is the acoustic invariance theory stronger due to use of static information?
It proposes that dynamic information is not necessary-invariance be specified with acoustic info from only one point in time
What are the additions of this theory to constructs already used to describe vowel perception?
1) Spectral analysis is only directed at a particular portion of the signal 2) There is a classification of the resulting spectrum by means of specific attributes or features
Stevens and Blumstein found that their classifications for place of artic were ___% accurate for ___.
85; initial stop consonants
However, they were less accurate for ____.
Stop consonants in the syllable final position
Do Stevens and Blumstein discount the contribution of formant transitions?
No; they were just looking for the minimal amount of info needed to determine place of articulation
What was Repp's theory for integrating formant transitions and the shape of the spectrum at consonantal release?
The release of the consonant can provide a static "snapshot" while formant transitions can provide more dynamic information. Each can be used to different degrees depending on the situation.
What is dichotic listening?
When different, competing sounds are presented to the two ears
Which ear does a better job at reporting verbal stimuli?
The right ear-because the left hemisphere is better at processing verbal stimuli
____ connections are stronger than ___ connections in the brain.