Speech science - models and theories

model

simplification of a system and its partsrepresent the system to be manipulated in a controlled manner

types of models

mechanicalphysiologicalmathematical/computer

theories

statement about a phenomenon incorporating underlying principles and assumptionsexplain observed datamake predictionsalways changing, never "proved

issue related to speech motor control

1) serial order issues2) degrees of freedom3) context sensitivity

spoonerisms

speech error in which the serial order of a word is produced in the wrong sequence

speech degrees of freedom

each potential muscle contraction (lots of muscles)respiration, larynx, articulationmovement in dif ways, speeds, and combinations

context sensitivity

speech sounds are produced differently based on their linguistic context

context sensitivity is influenced by

speaking ratestressclarity of articulation

context sensitivity

coarticulation is a type of...

dynamic systems model

model that reduces degrees of freedomexplains how muscles link together to perform tasks as a groupthese groups can change

connectionist model types

spreading activation modelsparallel distributed processing models

connectionist models (spreading activation, parallel distributed processing)

non-hierarchal processing, parallel info flows in all directionsspeech is a number of connected nodesexcitatory and inhibitory signalscoarticulation

target model types

spatial/articulatoryacoustic/auditory

target models

models that describe physical specification of a planned phonological unitdescribe sounds as a potential end result

spatial/articulatory target model

focus on position of articulatorsinternalized map of vocal tracts

acoustic/auditory target model

focus on acoustic output

when do we rely on feedback

learning speech soundssuprasegmentals

when do we rely on feedforward

once speech is developed

feedback steps

1) error is detected2) neural command for correction3) transmission of command to muscles4) contraction of muscle for the correction

feedforward

adjustments so the system is primed to move in efficient and coordinated manneruses previously learned commandspredictions of sensory consequences allow for quick estimations

Directions into Velocities of Articulators (DIVA)

incorporates feedback and feedforward