SPA 3011: Introduction to Speech Science

Notes:

Date	Topic	Reading	Assignment
T, 6/26	Coarticulation	BHR (5, 134-151)	Quiz 4

1. Coarticulation
   1. The production of speech sounds is affected by context
      1. In producing speech, the articulators are traversing through a rapid sequence of articulatory targets
      2. It is often the case that the articulators don't reach the same point they would for a speech sound produced in isolation
   2. Types of coarticulation
      1. Assimilation
         1. In assimilation, adjacent speech sounds that require the same articulator use a single articulatory gesture for both sounds
         2. For example, in the phrase "I miss you", the /s/ and /j/ phonemes both require particular articulations of the tongue tip and blade
         3. In this case, the /s/ is often produced with the palatal gesture of the /j/, resulting in a "sh" sound (and similar sequences are found with other alveolar and palatal combinations, like in "did you"
         4. Other examples: /t/ and "th" sequences produced with a single dental gesture, /n/ and /k/ or /g/ sequences produced with a single tongue gesture (compare "unbelievable" to "uncontrollable"), velar consonants before front versus back vowels (compare "key" to "cool")
      2. Coproduction
         1. In coproduction, adjacent speech sounds that use different articulators can be overlapped in production
         2. For example, the /s/ does not require the use of the lips, so lip rounding in an adjacent sound (like /u/) can begin during the /s/ (compare "seat" and "suit", similarly for "tea" and "two")
         3. In general, there is always some coproduction between adjacent sounds (e.g. the closure for the /t/ in "active" will take place before the release of the closure for the /k/, resulting in just one stop burst)
         4. Coproduction of consonants and vowels provides additional information about the identity of both sounds in the formant transitions between the sounds
      3. Hypoarticulation
         1. Variation in the degree to which the articulators reach their "ideal articulatory goals" is referred to as degrees of hyperarticulation (very careful pronunciation) and hypoarticulation (pronunciation that undershoots the target)
         2. In rapid speech, the corner vowels often do not achieve the extreme articulations that they would if produced in isolation
         3. Similarly, the central vowel (schwa) does often achieve the fully neutral position, and is a little front, back, high, or low depending on the surrounding vowels and consonants

2. Spectrogram reading
   1. Determining what was said in speech just by looking at a spectrogram is extremely difficult
   2. Some gross patterns are easy to identify
      1. Formants for vowels and semi-vowels
         1. Front vowels have large gap between F1 and F2 (high F2)
         2. Back vowels have small gap between F1 and F2 (low F2)
      2. Low F3 for /r/
      3. Formants of lower amplitude for /l/ and nasals
      4. Long duration frication noise for fricatives and affricates
         1. Loud fricatives at high frequencies are alveolars and palatals
         2. Quiet fricatives over a broader frequency are labio-dentals and interdentals
      5. Period of silence, release burst, and formant transition for stops
   3. These gross patterns are usually sufficient to align a spectrogram with a transcript of what was said

3. Feedback in speech
   1. Talkers monitor their speech using feedback from several sources
   2. Auditory feedback
      1. For listeners with adequate hearing, the sounds of speech can be monitored through the auditory system
      2. The relevance of auditory feedback has been demonstrated in experiments using delayed auditory feedback
         1. In a delayed auditory feedback experiment, the participant listens to their own speech production through headphones
         2. Speech becomes disfluent if auditory feedback is given at a delay of 200 ms
         3. Children also become disfluent with delayed auditory feedback, but the delay depends on age, with younger children requiring a longer delay
      3. Additional evidence comes from experiments that modify the feedback (for example, removing information in some frequency range) and find that speakers change their behavior to compensate for the change
      4. More obvious examples come from everyday uses of speech as well, such as speaking up when talking in a loud environment
   3. Tactile feedback
      1. Feedback also comes from the contact of articulators with points in the vocal tract
      2. The importance of tactile feedback has been studied using bite blocks and other means of interfering with articulation or feedback (like blocking nerves)
      3. Other articulators will compensate for the block to make sure the articulatory goal is reached
      4. Compensation does not occur solely on the basis of reaching an articulatory goal, as the acoustic result of that goal can also be monitored as auditory feedback
   4. Proprioceptive feedback
      1. The articulators also have sensory information about their position and orientation
      2. Most research on proprioceptive feedback comes from domains other than speech
         1. In the absence of visual or tactile feedback, people are aware of the position of their limbs and body
         2. People can use proprioceptive feedback to accurately judge the size, weight, and configuration of items they are holding
      3. We assume that these results are applicable to speech as well