Contrasting Oral Gestures

 

What allows oral constiction gestures to be systematically distinguished from one another?

(1) Anatomically distinct constrictors

Shared by all members of the community

Decomposition of body into distinct organs is at least partially innate.

These contrast in all languages.

Distinct constrictions emerge in early babbling of infants.

Alternative theory

MacNeilage and Davis theory: Infants only oscillate their jaws, and depending on the posture of their tongues, they produce what adults hear as or [de] or [gu].

See example

 

(2) Constriction degree

potential continuum
how partitioned into discrete regions that are systematically distinguished across speakers?

Quantal Theory of Speech (Stevens, 1968, 1989)

Examine relation between:
dimensions of physical device and sound produced

Experiment with a straw

  1. Blow gently into straw, so it makes only soft noise.
  2. Start gradually narrowing just behind tip of straw.
  3. At some point, sound will suddenly become loud and/or high pitched.
  4. Loud sound will remain until straw is tightly pinched.

 

Results of experiment: map relating opening and sound

 

There are stable and transition regions in the map.

Stable regions
Small change in opening do not effect sound

Transition regions
Small change in opening shifts from one stable region to another.

 

Stevens' Quantal Theory

Stable and transition regions generally characterize relations between dimensions of sound producing devices and the sounds they produce.

This is true of the human vocal tract.

The nonlinear nature of this map partitions a potential gestural continuum into distinct stable regions.

Languages employ contrastive gestures that are produced in distinct stable regions of the map relating articulation and sound. why?

 

Example: Constriction Degree

 

Interaction of Constriction Degree and Voicing

 

Demonstration:

  1. Produce [v]
  2. Widen constriction until turbulence disappears.
  3. Now stop voicing
  4. Turbulence is back, why?

Turbulent flow of a liquid (or gas) involves generation of random (or chaotic) patterns of molecular vibration.

It depends on:

 

Reynolds Number (Re) is a dimensionless quantity that takes both into account.

A given channel area (CD, constriction degree) will have different Reynold's number (Re), depending on the volume velocity of the flow.

Turbulent flow will result at a particular Re threshold (for speech conditions, Re=1700), and the intensity of the generated turbulence will increase as Reynolds number increases above threshold.

see graph

 

In general, voicing reduces volume velocity (glottis is closed roughly half the time).

CD<20 mm2

Re>1700

20<CD<100 mm2

Re> 1700 for voiceless flow rates
Re< 1700 for voiced flow rates

CD>100 mm2

Re<1700

 

Effects of voicing and turbulence interaction on gestural patterning in languages

statistical rarity of voiced fricatives

Aerodynamic conditions for voiced fricatives

Voicing: subglottal pressure > oral pressure

Turbulence: oral pressure > outside pressure

What happens to them?

  • devoicing
    to preserve turbulence, voicing may be lost
  • loss of turbulence
    to preserve voicing, constriction degree may increase

CD can be partitioned into four regions (Catford, 1977):

stop

no flow

fricative

turbulent flow

approximant

turbulent if voiceless

resonant

laminar flow

 

What is the gesture's goal?

 

Ladefoged Classification

Lateral vs. central fricatives

Zulu

Trills

aerodynamic phenomenon

depends on:

  • airflow
  • constriction degree
  • articulator tension

     

Constrictors that can trill:

Constriction Degree of trills

intermediate between stop and fricative

articulators touch, but are not compressed: CD = ~0

 

Summary of constriction degrees (from Catford, 1977):

 

 

 

 

 

Top of Page | Linguistics 120 Home page