speaking Flashcards
how is speech produced ?
we create sound waves by taking quick breaths of air which are steadily released by the intercostal muscles in the rib cage
the passage of sound waves is aided or impeded by opening or closing different speech articulators (tongue, velum, teeth and lips)
voice pitch is controlled by the length of the vocal tract
the folds of the larynx can be stretched over the glottis to restrict air flow (voicing)
the tension and position of vocal folds change the pitch of a sound
chambers in the vocal tract: pharynx, oral cavity, lips, nose; these change
the stream of sound is to produce vowels and consonants
how does the vocal tract act as a filter?
Voicing creates a sound with a dominant frequency (pitch) and many harmonics
the chambers above the larynx each allos some frequencies to flow through while filtering others
speech production involves altering the shape and size of these chambers to alter the resonant frequencies (formants) produced
what are sound waves
sine waves consisting of pitch (frequency: cycles/sec) and loudness (amplitude: height) combine to create a complex waveform
what are the properties of spectrograms
sound pressure time plots
x = time
y = frequency
dark bands: formants - concentrations of acoustic energy in a frequency bands
the relative spacing of formants produces percepts of different vowels
steady state formants are constant
formant transitions change over time
F1 = first formant
Increasing F1 corresponds to lower vowels
F1 corresponds to openness of vocal tract (lower tongue)
F2= second formant
F2 represents roundness of vowels (further back tongue)
High F2 corresponds to tongue being more forward
how are vowels pronounced
F1 increasing down page (y) F2 increasing across page (x)
relates to position of tongue
steady state formant
open orla tract from glottis to lips no other articulators moving
how are consonants formed
blocking the passage of air either partially or completely plosive/ fricative place of articulation is varied bilabial (lips) alveolar (longue behind teeth) velar (back of vocal tract)
how has the vocal tract evolved for speech
shorter, rounder tongue
larynx lower in the throat for greater movement of tongue
(computer modelling suggests monkey larynx can produce speech sounds - it is the brain mechanisms that they lack)
what is stage 1 in lemma model (Levelt, 1989,2001)
lexical selection
1. conceptual focusing and perspective taking (as this is an act of communication we must consider what concepts we want to convey and to whom)
we then need to map these onto the lemma
2. lexical concepts: units that integrate all the sematnic features associated with the word
3. lemma selection
then we arrive at the lemma: an abstract representation of a word that integrates, meaning, sound and grammatical associations (grammar is particularly important in languages where words are gendered, for example)
what is stage 2 of the lemma model
form encoding
1. retrieving morphemic phonological codes (how the word is going to be said)
2. phonological codes (consonants/vowels)
3. prosodification and syllabification (where will the stresses in the word be/ tone if multiple languages spoken)
4. phonological word
5. phonetic encoding (motor codes)
arrive at the articulatory score
what are the key features of the lemma model
two separate stages that are qualitatively different
in lexical selection there is lots of competition within LTM as many lexical representations with different weights (based on experience) are activated
in contrast form encoding is 1:1 mapping of a selected lemma onto a motor code
in a meta-analysis of speech production studies indefrey & levelt (2004) found what?
they were able to identify the areas associted with the linguisitc network
while there was some modality specifc activatoin there was co-activation of core, speech specific areas: anterior inferior frontal lobe and superior temporal gyrus (lexical retrieval and phonological areas and perceiving produced speech) as well as the (pre)supplementary motor area(s) (form encoding)
what are the problems with the lemma model
modern models support concurrent parallel processing of incoming speech perception and production
there are examples of patients with modality specific deficits which are not explained by this model (e.g patient able to write verbs but not nouns, but can speak both, Caramazza 1997)
evidence for interaction between different steps of models (e.g mistakkes in speech production that appear similar (e.g cat/rat vs rat/sat) suggest mistakes are occuring in lexical selection not just form encoding
what is the importance of speech perception
for guiding speech production
involves left lateralised dorsal stream consisting of posterior temporal, parietal and frontal regions
what does (Tang et al., 2019) altered auditory feedback study show
Experiment: Participant says “head”,
F1 is shifted up and feedback in real time. Sounds closer to “had”.
Participants alter production to compensate for feedback by lowering F1and increasing F2 to produce something closer to “hid”.
what does Buchsbaum et al’s2011 lesion overlap study show
14 patients with conduction aphasia
Posterior sylvian fissure to cortex
Similar lesion with white matter
Rehearsal of WM indicates phonological loop
Sylvian parietotemporal critical for mapping of phonology to articulatory loop for production
In the dorsal stream
Sensory-motor interface
