This web page presents an introduction to the gesture in Articulatory Phonology and interactive apps to play around with some of the parameters of the gesture.
In Articulatory Phonology, an articulatory gesture is seen as the basic building block of phonology. Multiple gestures are organized in gestural scores to form words. It is important to note that these gestural scores are seen as phonologoical. This means lexical entries are represented as structures of gestures rather than as string of segments. This image shows the gestural scores for "mad" and "ban". Each block represents the interval in which a gesture is active:
[Gestural scores adapted from Goldstein, Louis, Hosung Nam, Elliot L. Saltzman & Ioana Chitoran. 2009. Coupled oscillator planning model of speech timing and syllable structure. In G. Fant, H. Fujisaki & J. Shen (eds.), Frontiers in phonetics and speech science, 239–250. The Commercial Press Beijing.]
Both words have an alveolar closure in the tongue tip and a closure of the lips. They also have the same vowel gesture in the tongue body. The timing of velum opening (velum: wide) makes the difference between the two words, as it comes early in "mad" and late in "ban".
Each gesture is modelled as a dynamical system. For now, it is enough to understand that a dynamical system is a system that changes its state over time. Articulatory phonology adapted the idea of a mass-spring system as the basis for the gesture. Mass-spring systems are well-studied dynamical systems in physics and they are able to model a variety of phenomena. An animation of such a system is shown in the next image:
[Animation from wikimedia commons, user Svjo.]
In this system, a mass is attached to a spring. When you pull on the mass, the spring will be stretched. When you release the mass, the mass will be pulled back by the spring. Usually, we observe that a mass-spring system oscillates for some time until it eventually comes to rest. This is, of course, not what we observe with articulatory movements! Our articulators do not oscillate. Instead, they go immediately to a specific location. For example, the tongue moves to a constriction location at the alveolar ridge.
To model this behavior, the mass-spring system used in Articulatory Phonology is critically damped. Damping occurs everywhere in the world around us. A pendulum, for instance, swings for some time during which the amplitudes of the movement increase. Eventually it comes to a halt. In the case of critical damping, the damping is even stronger. This means that the mass will not oscillate but instead go immediately to its resting position. This resting position is often called attractor or target of the gesture.
The next image illustrates this behavior. It shows the position of a gesture over time. We could imagine this gesture representing the lip opening movement as we produce the syllable [ba].
Gestures can have different targets as shown in this illustration:
We will not go into the details of the mathematics that define the gesture but we will explore some of the factors that shape the gesture. One of these factors we will look at is the duration of the interval in which an a gesture is active. The other factor is the target of the gesture, i.e., the attractor that the gesture converges to. In the following sections, you will be able to try out the consequences of changing these parameters yourself.
The first factor we consider is the duration of the activation interval. From the gestural scores you know that gestures are active for a period of time. In this section, we explore the effect of the duration of the activation interval on the gesture.
We have two lip opening gestures, red and grey. Each opening gesture is followed by a closing gesture. In the top panel, you see the gestural activations. You manipulate the red opening gesture. You can compare the change in this gesture to the grey gesture which is the reference gesture. More specifically, you change when the gesture ends.
Below the gestural activation plots, you find the position of the gesture and the velocity. In the velocity panel, you see one peak for each gesture. The earlier peak corresponds to the opening gesture. The later peak corresponds to the closing gesture. You'll notice how the peak for the red closing gestures moves to the left (earlier point in time) as you shorten the activation of the opening gesture. This is because the closing gesture comes earlier when the activation duration of the opening gesture is shorter. Focus on the first peak, the velocity peak of the opening gesture.
Question: What happens to the position of the gesture when the activation interval is shorter?
Question: What happens to the velocity when the activation is shorter?
Question: What happens to the duration from the onset to the lowest position when the activation interval is shorter?
We have seen in the introduction that gestures can have different targets. In this section, we explore the consequences of changing the gestural target.
The following app lets you manipulate the target of a gesture. We have two lip opening gestures, red and grey. Both gestures are followed by lip closing gestures. In the top panel, you see the gestural activations. In this example, they will not change. Below the activations of the gestures you find the position over time and the velocity. In the velocity panel, you see one peak for each gesture. The earlier peak corresponds to the opening gesture. Observe what happens to the velocity when the target changes.
Question: What do you observe about the velocity when the target is more extreme (here: lower)?
Question: What happens to the duration between gesture onset and lowest point when the target is manipulated?
Question: Why does the velocity increase?
Let's briefly summarize what we learned about the gesture in general:
What did we learn about the activation interval?
What did we learn about the target?