Phonetics in Conversation

Transcription

Phonetics in Conversation
Phonetics in conversation
Paper 9
Foundations of Speech Communication
Michaelmas Term: Week 2 (practical class)
Katharine Barden
Aims
• To think about what is conveyed during
speech communication
• To introduce the anatomy and physiology
of the larynx and its role in communication
• Introduce examples from Conversation
Analysis
Communication
• Usually involves more than one person
(interaction)
• A means ‘to get something’
– for babies: food, attention, comfort
(maintain biological equilibrium)
– for older people: the same, though we
normally think of them as more complex e.g.
“understanding, friendship, discussing ideas,
making decisions and contracts”
Linguistic units
• Traditionally, focus on contrastive lexical
meaning (identifying words)
• Emphasis on:
– Phonemes: abstract units related to the sounds that
distinguish words
• e.g. sap vs tap contrast in terms of the phonemes /s/ and /t/
– Allophones: physical phonetic differences, more or
less predictable from word structure, that, if done
“wrong”, do not change the meaning of the word but
could make its pronunciation sound strange
• e.g. ‘clear’ vs ‘dark’ /l/ in SSBE: [l] (syllable-initial) vs [ɫ]
(syllable-final) - lull [lʌɫ]
– Features: abstract dimensions that relate to the
sounds that distinguish phonemes
• e.g. /p/ and /b/ contrast in the feature ‘voice’
Broad approach to communication
• But, within an interaction, a lot more is
conveyed than linguistic units
• Phonetic cues can contribute to linguistic,
paralinguistic and extralinguistic aspects of
communication
• ‘Meaning’ derived not just from what is in
the speech signal, but how it relates to
prior experience
The larynx
Biological function:
a valve to protect the lungs
The vocal apparatus
Source:
Lieberman and
Blumstein,
Speech
Physiology,
Speech
Perception and
Acoustic
Phonetics
(1988).
3 main parts: subglottal system, larynx, supralaryngeal vocal tract
Laryngeal anatomy basic checklist
1. 4 main cartilages (cricoid, thyroid, pair of arytenoids, epiglottis)
•
joined to each other and slung from one bone (the hyoid) by
membranes
•
joined to bones by extrinsic muscles – these fix it or move it in the
neck
•
joined to each other by (mainly paired) intrinsic muscles which
– move the cartilages relative to one another (4 main pairs)
– comprise the bulk of the vocal folds (2 pairs)
2. The vocal folds are inside (and thus part of) the larynx
•
bundles of muscle, ligament and mucous membrane
•
extend horizontally from the front (thyroid notch) to the back
(arytenoids)
•
space between them is the glottis
3. Laryngeal musculature enables vocal fold closure and opening
(affecting size and shape of the glottis) , and all adjustments for
phonation
The larynx
anterior (front) view
The larynx
posterior (back) view
Structure of the larynx
• 3 +1 main cartilages:
– large thyroid (Adam’s apple) comprising 2 plates and 4 horns.
connected upwards to hyoid bone by thyrohyoid muscle/ligament)
– smaller, circular cricoid with ‘signet ring’ shape: higher at back
than front
– 2 small, pyramid-shaped arytenoids sitting on top of posterior
surface of cricoid
– (+ epiglottis: up from thyroid angle, rests against back of tongue)
• Vocal folds connect vocal process of arytenoids to inner
front of thyroid cartilage
Front view
Rear view
Side view
View from top
Inside the larynx: the vocal folds
mid-sagittal view
Glottis = space between folds
Vocal folds can be in an open
(abducted) or closed (adducted)
configuration
View from above:
Folds open (abducted)
View from above:
Folds closed (adducted)
Phonation (voicing) basic checklist
1. To phonate, the vocal folds must vibrate
2. To vibrate, they must be held close
enough together to impede the airflow
through the glottis
3. Muscles bring them together & hold them
there
4. The transglottal airflow itself sets them
into vibration, and maintains the vibration
– myoelastic aerodynamic theory of phonation
(elastic recoil and Bernoulli forces)
Vibration of the vocal folds
results in phonation (voicing)
Myoelastic aerodynamic theory of vocal fold vibration
(van den Berg, 1950s)
1. Muscular activity rotates and rocks the arytenoid cartilages
so that their vocal processes come together in the midline,
thus positioning the vocal folds close together or in actual
contact.
2. Air pressure increases below the glottis until folds forced
apart. (The subglottal pressure increase leads to a
transglottal pressure drop.)
3. Air travels faster through the glottis when it is narrow. This
causes a local drop in air pressure (Bernoulli effect) which
causes the folds to be sucked towards each other.
4. The Bernoulli effect, together with the elastic recoil force
exerted by the displaced vocal folds, causes complete
glottal closure again.
5. The process begins again at step 2.
Vocal folds
during a vibratory cycle
Open for
breathing
http://cspeech.ucd.ie/~fred/teaching/oldcourses/phonetics/pics/vfold1.gif
Vertical views of the vocal folds
during one vibratory cycle
The folds are threedimensional, and they vibrate
in three dimensions.
1
4
The pattern of vibration is like
a ‘wave’ travelling up them.
The lower sections part first,
and come together first.
2
5
‘Cover’ (outer layer) and ‘body’
(inner layers) of folds are often
distinguished because they
vibrate fairly independently
3
6
After Stevens (1998) Acoustic Phonetics
(Baer, 1975)
Controlling phonation:
Intrinsic laryngeal muscles
This lecture does not address
external laryngeal muscles, nor
detailed vocal fold anatomy
(read e.g. Hardcastle)
No phonation, or stopping phonation
Front view
Rear view
Side view
from above
• Abduction: Vocal processes of arytenoids
(front part) rotated backwards and
outwards (posterior cricoarytenoid
muscle)
• This moves the vocal folds apart and so
widens the glottis
Starting and maintaining phonation
Front view
Rear view
Side view
from above
• Adduction: vocal processes of arytenoids
moved together (lateral cricoarytenoid,
interarytenoid muscles)
• This brings the vocal folds together, thus
closing the glottis
Coordinating glottal and oral constrictions
oral closure
oral constriction area
oral release
stop
VOT
glottal constriction and vibration
[aba]
voiced
negative
[apa]
voiceless unasp.
zero
[apha]
positive
[ahpa]
voiceless
aspirated
preaspirated
[aʔpa]
glottalised
“zero”
[abʱa]
breathy
time
key
top row: complete oral closure; all other rows: vocal folds adducted but not vibrating
top row: oral articulators open; all other rows: vocal folds abducted and not vibrating
modal phonation: vocal folds adducted and vibrating
breathy phonation: vocal folds partially adducted and vibrating
Periodic sounds
Fundamental frequency
(f0) = 1/period
period ≈ 6.3 ms
F0 ≈ 160 Hz
period ≈ 5.8 ms
F0 ≈ 172 Hz
period ≈ 11.2 ms
F0 ≈ 90 Hz
shows 1 period
Pitch control
Front view
Rear view
Side view
from above
• Increasing pitch: contracting cricothyroid muscle: pulls
front of cricoid up towards thyroid, so back of cricoid
moves down and back, taking arytenoids with it and
stretching/tensing vfs → vibrate faster
• vocalis – complex adjustments to length, tension,
thickness & medial compression of vocal folds
Pitch control
Front view
Rear view
Side view
from above
• Increasing pitch: contracting cricothyroid muscle: pulls
front of cricoid up towards thyroid, so back of cricoid moves
down and back, taking arytenoids with it and
stretching/tensing vfs → vibrate faster
• vocalis – complex adjustments to length, tension,
thickness & medial compression of vocal folds
Voice qualities
Classification systems vary from very simple e.g.
creak
- modal
- breathy,
Ladefoged (2001)
Vowels and consonants
to very complex e.g. Laver
Open quotient = the proportion of the cycle for which the
vocal folds are open
Low for creaky voice
Approx 0.5 for modal voice
High for breathy voice
Communicative uses of voice quality
• Indexical: part of an individual’s
characteristic speech patterns
• Cultural: some cultures have distinctive
voice qualities
• Communicative function:
– controlling conversation cf. ‘so’, ‘I think’, ‘and’
– conveying affect (emotion)
• Phonemic contrast
Pathological disorders
of vocal fold vibration
•
neural: e.g.
– paralysis, spasmodic dysphonia → incomplete closure → breathy →
quiet; usually high pitch; or harsh if tense.
– Parkinson’s → immobile + tremor, quiet, restricted pitch range (often
high), hoarse
• viral laryngitis → inflamed VFs → hoarse, or silent
• habitual abuse (shouting, smoking) → hoarse, harsh
• physical damage to the folds (nodules, scars....): incomplete closure +
irregular vibration → breathy, hoarse, low volume
Vocal fold nodules
Sulcus vocalis (vocal fold scar)
Conversation analysis
• Developed by sociologists in early 1970s
• Emphasis on the structure of interactions
Stand-alone “so” in American
English
holding-‘so’
the same speaker continues with more
on-topic talk
trailoff-‘so’
there may be a change in who talks
For more details, see Local (2007) ICPhS http://www.icphs2007.de/
(click on Plenary Speakers)
Phonetic characteristics
• In contrast to ‘holding so’, ‘trail-off so’
tends to be:
– quieter and lower in pitch than the accented
part of the final foot of the preceding speech
– it never has final glottal closure
– can be creaky throughout the whole syllable.
Turn co-completion
– once those cameras start flashing
particularly with the infants it puts them off
– but when we walk out of the class nobody
knows what went on
• There is nothing remarkable about these
stretches of coherent speech – except that
they are jointly accomplished by two
speakers rather than one
Turn co-completion
(1) C&M Xmas Photo
K: once those cameras start flashing
particularly with the infants .hh
C: it puts them off
K: it puts them of[f and
C:
[yeh
Courtesy John Local
Turn co-completion
(2) Two Girls (5)
B: .hhh and we nod when he wants us to say yes (h)e[n ]
.hhh
A:
[ye]ah
B: we raise our hands when he wants to take
a poll [.hh you know
A:
[mm
B: but when we walk out of the class
A: nobody knows what went on
Courtesy John Local
Talk which completes another’s
turn is:
1.
2.
3.
4.
not slower than the talk they complete
not louder than the talk they complete
rhythmically integrated with prior talk
narrower in pitch range than the talk they
complete
5. closely pitch-integrated with prior talk
Turn co-completion
Courtesy John Local
Summary
• Speech communication is about more than
conveying linguistic units
• Phonetic cues can convey many other
types of meaning which are crucial for
successful interactions