Models And Theories Of Speech Production

Front 1

What is the definition of a theory?
(concept)

Back 1

> a statement about a phenomenon incorporating underlining facts, principles, and assumptions.
-subject to change by new/modified theories

Front 2

What is the definition of a model?
(explain theory)

Back 2

>a simplification of a system that can be manipulated in a controlled manner
>it is the physical, physiological, or mechanical version of the system to be tested.

Front 3

Physiological models use what specimen?

Back 3

use specimens taken from animal or human cadevers to determine how a particular structure responds under different conditions

Front 4

source filter theory tested how?

Back 4

take a pipe close one end and open the other and you can describe how the formant frequency changes.
F=1/m
use pendelum model

Front 5

speech production and perception theories are?

Back 5

>complex
-contain subsystems

Front 6

name some subsystems

Back 6

>auditory
>nervous
>articulatory
>respiratory
>phonatory

Front 7

so what is best due to complexity?

Back 7

have theories that integrate current knowledge and models that test how they work and breakdown

Front 8

what do you need to know about break downs?

Back 8

how they can be compensated and repaired

Front 9

three major issues related to organization of speech production

Back 9

1. the serial order issue
2. degrees of freedom
3. context sensativity problem

Front 10

serial order issue

Back 10

>speech is a sequence of elements
example: /k/ /ae/ /t/ can be used for cat, tack, and act
>elements can be specific features of a sound, phonemes, syllables, parts of syllables, etc.
dosnt have same meaning though

Front 11

degrees of freedom

Back 11

>movement of one structure(muscles, of respiratory, laryngeal, and articulatory subsystems) of a subsystem in relation to another
(lower lip and jaw can but also potentially cant be in phase with each other)

>the speech motor system must regulate all the muscular contractions (rapid movement and correct sequences) of all the speech suvsystems to generate the target sounds and words.

Front 12

degree of freedom definition

Back 12

each different muscular contraction in a system is known as degree of freedom
-this is huge in speech production
-n= infinate all degrees of freedom

Front 13

context-sensativity problem

Back 13

>speech is influenced by speaking rate, stress, clarity of articulation, and other factors
>one speech sound can be produced in several ways- this variablity is a central factor of speech motor regualtion

example: coarticulation is an integral aspect of apeech production that results in variability in production of the target sound.

Front 14

theories of speech production

Back 14

1. target models
2. feedback models
3. feedforward models
4. dynamic systems model
5. connectionist model

Front 15

target models

Back 15

target models describe speech as " a process in which the speaker attempts to attain a sequence of targets corresponding to the speech sounds he is attempting to produce"

Front 16

about targets

Back 16

targets are spatial. there is an internalized map of the vocal tract in the brain that allows the speaker to move his/her articulator
-movements for specific phoneme must change depending on starting point
example: to achieve /k/ the tongue moves in a different trajectory depending on the preceding vowel (/a/,/u/, or /i/).

Front 17

about targets

Back 17

they are acoustic-auditory-
>goal is the acoustic output but the articulatory movements used to achieve these acoustic output may vary (depending on factors such as adjacent sounds, rate of speech, and stress patterns in speech.

Front 18

summary of target models

Back 18

they describe speech as a series of either spatial or acoustic auditory targets that a speaker attempts to achieve.

Front 19

target model process

Back 19

brain (internalized map of the vocal tract) when you speak it is forming then coming out ------> spatial targets--------> articulators (activated and speech produced)--------->auditory and somatosensory fedback to regulate fine movements and corrent any errors.---------> once produced sent back to brain (determines whether produced correctly or not)

Front 20

feedback model

Back 20

refers to the transfer of part of the output of a system back to the input to regulate correct any errors in the output

output=articulatory or acoustic signal
feedback= auditory, proprioceptive, kinesthetic and tactile, sensory receptors

Front 21

feedback process

Back 21

central processor----> if actual and intended output not similar (error signal created)------>vocal tract(muscles)------>speech (the actual output)--------> auditory/proprioceptive/tactile/kinesthetic------->internal model(intended output)

Front 22

feedforward process

Back 22

central processor-----> vocal tract (muscles) --------> speech (output)------> self monitoring loops back to vocal tract (muscles)

-very fast process because does not go to brain goes to the vocal tract

Front 23

feedback vs. feedforward

Back 23

feedback-
>used to detect and correct errors in speech output

>signals from the periphercy (vocal tract) to a central processor in the brain for comparison of tje intended and actual movements

-it is a slow process

Front 24

feedback vs. feedforward

Back 24

feedfoward-

>used to make articulatory adjustment online
>signals adjusted in the periphery (vocal tract) so that the system is primed to move in an efficent coordinated manner.

-fast process

Front 25

dynamic systems model

Back 25

speech as a dynamic pattern of trajectories through articulatory adjustments

trajectores=movement, direction, speeds, etc.

Front 26

dynamic systems continued

Back 26

>groups of muscles link up together to perform a particular task
this is called synergy or a coordinative structure
>these sets of muscles behave as an integrated unit
>different muscular responses in a synergy can be adjusted to meet the requirments
ex: lip and jaw coordnative in bilabial closure

Front 27

speech perception problems

Back 27

1.linearity and segmentation
2.speaker normalization
3. basic unit of perception

Front 28

linearity

Back 28

asserts that a specific sound in a word corresponds to a specific phoneme and all of these occur in a particular sequence

Front 29

segmentation

Back 29

is based on the notion that the speech signal can be divided into discrete units that correspond to specific phonemes.
•Speech perception is based on a linear correspondence between the acoustic signal and the linguistic phonemic units

Front 30

example linearity and segmentation

Back 30

Ex: /k/ in /ku/ is different from /k/ in /ki/
•But the acoustic boundaries are blurred

Front 31

speaker normalization

Back 31

•(Speakers vary across age and sex; also vary in speech sounds in terms of pitch, loudness, stress and rate of speech) Ex: formant frequencies vary from speaker to speaker, still most listeners do not have difficulty in vowel recognition.
•So, how do listeners recognize speech sounds and words despite large variations in the way speakers produce them ?
•Because we normalize speech; that is we are able to ignore irrelevant differences between productions of a given sound while focusing on the features that characterize it as a member of that particular phoneme family.

Front 32

basic unit of perception

Back 32

what kind of units are more salient for a listener ? Do we perceive acoustic-phonetic features, allophones, phonemes, syllables or larger units of speech ?
•Individuals rely more on one kind of unit than another depending on the situation and the context of the speech to perceived.
•Ex: quite listening environment (pay attention to smaller units) vs back-ground noisy environment (pay attention to higher level of linguistic environment)

Front 33

infants and young children

Back 33

Infants and young children (may process auditory information using larger units such as syllables) vs. older children and adults (may rely on smaller units such as phonemes)
•Children- go through a developmental weighting shift- perceptual strategies change with increasing linguistic experience

Front 34

theories of speech perception

Back 34

Motor theory
•Acoustic Invariance theory
•Direct Realism
•TRACE Model
•Cohort theory
•Fuzzy Logical Model of Perception
•Logogen theory
•Native Language Magnet theory

Front 35

categories of speech perception

Back 35

Active vs. Passive
•Bottom-up vs. Top-down
•Autonomous vs. Interactive

Front 36

active theory

Back 36

an individual’s knowledge of speech production facilitates sound recognition (linking between speech perception and speech production)

Front 37

passive theory

Back 37

stress on filtering mechanism of the listener (emphasize on the sensory aspects of speech perception, minor role of speech production

Front 38

bottom-up

Back 38

All information necessary for recognition of sounds is contained within the acoustic signal

Front 39

top-down

Back 39

Higher level linguistic and cognitive operations play a crucial role in the identification and analysis of sounds

Front 40

interactive

Back 40

Information and knowledge from many sources are available to the listener at all stages of speech perception system

Front 41

autonomous

Back 41

Signal processed in a serial manner; output of one stage of processing provides the input to the next stage

Front 42

where was the motor theory developed??

Back 42

Developed at the Haskins Laboratory at the Yale University

Front 43

motor theory

Back 43

Emphasizes the link between speech production and speech perception in terms of articulatory gestures that individuals are innately able to perceive. (Because listeners have experience in producing speech sounds themselves, they are aware at some level of the relationship between movements of the articulators, vocal tract configurations and the acoustic consequences of these articulator movements and positions.)

Front 44

more about the motor theory

Back 44

Assumes that speech perception is unique, relying on a special processor located somewhere in the brain to decode speech
•Abstract gestures produce particular constrictions in the vocal tract, with each constriction being appropriate for a specific phonetic place and manner of articulation. These abstract articulatory plans are known as gestures.

Front 45

TRACE model

Back 45

Is a connectionist model that accounts for integration and parallel processing of multiple sources of information in speech perception
Assumption: Behavior can be modeled

Front 46

features

Back 46

Features:
1.Network of units are phonetic features, phonemes, word
2.Feedback and feedforward links between units
3.Connection between upper and lower units is bidirectional/ facilitatory (both top-down or bottom-up)
4.Connection within a level are inhibitory (target feature or word will be activated and competing features or words will be suppressed)