Dynamics of Speech Production

Coarticulation

 

Recall from the Speech Chain.....

Linguistic Level – we plan what we wish to communicate, and what linguistic units we will use to communicate.  What linguistic units do we plan for?

Our plan has to include strategies for how we “string these units together” because speech is temporally ordered.  That is, in speech the temporal order in which things are produced is important.

Syntax, for example, is the set of rules that govern word order.  Also, the order of units within words makes a difference, as demonstrated by the difference between /kQt/ and /tQk/

Our first inclination regarding the time ordering of speech elements might be that we order words (and morphemes) within a thought phrase, then order phonemic elements within each word.  Thus, the “units” we use in processing speech at the linguistic level would be words or syllables.  Our task, then, would be to 

We cannot directly examine what happens on the linguistic level of spoken language production.  However, we can examine how articulators move, which should reflect the underlying linguistic organization we use to generate spoken language.

When we examine articulation of speech, we see many things that tend not to support a view that we produce speech  in “units” of phonemes, syllables, or words, strung together in a temporal sequence.

Much of the contrary evidence is found in a phenomenon that has been called “coarticulation.”

Coarticulation is a phenomenon where a feature of a phonemic unit is achieved in the articulators well in advance of the time it is needed for that phonemic unit.

As a general rule, we will achieve the articulatory manifestation of any feature of a speech sound as early in the temporal sequence as the rules of phonology will allow, often well before we execute other articulatory features of a phoneme.

 Example 1 – Coarticilation Within a syllable:

 Contrast these two CN monosyllables: /si/ - /su

Notice that the two syllabic nuclei differ from each other on tongue advancement.  Both are high vowels, /i/ is front and /u/ is back.  They also differ from each other on lip rounding.  The vowel /u/ has mandatory lip rounding while the vowel /i/ is neutral for rounding – it may be rounded or unrounded.  In English, lip rounding for front vowels is optional.  (Note that other languages have front rounded vowels so that in French, for example, front vowel lip rounding is not optional.)

In English, it makes no difference if we produce /i/ rounded or unrounded.  So, according to the general rule, we will achieve this articulatory manifestation (lip rounding) as early in the temporal sequence as we can.  In the case of the syllable, /su/, lip rounding is achieved while we are producing the sibilant, /s/.

Say the two syllables in turn, and pay attention to the articulation of lip rounding.  You will notice that in the syllable, /su/, you initiate lip rounding while you are producing the sibilant, /s/.  In addition, if you hear just the sibilant /s/ produced in these two contexts, but without the vowels, you can easily identify which /s/ was produced in the context of a back vowel as opposed to a front vowel.

The unavoidable conclusion is that we do not produce indivudual phonemic units.  We know this because we observe that phonemic units are articulated differently depending on context.

We would have guessed that /si/ and /su/ were organized in our linguistic systems as in the table below.  Notice that we would have predicted identical voiceless lingua-alveolar fricatives in the two syllables.

/CN/

First phonemic unit in sequence

Second phonemic unit in sequience

/si/

voiceless lingua-alveolar sibilant

high front vowel

/su/

voiceless lingua-alveolar sibilant

high back rounded vowel

What we observe in real human speech, however, is two systematically different sibilant productions, with the latter being consistently rounded in the context of a rounded vowel.

Apparently, phonemic units are not the basic building blocks of spoken language.

This is our first real example of Anticipatory Coarticulation.

Other examples:

  1. We see coarticulation across syllable boundaries so that we can conclude that syllables are not the basic building blocks of spoken language.

Consider /k«nstrIkt/ (“constrict)”as contrasted with /k«nstru/ (“construe”). Lip rounding in the the word, “construe,” begins as early as the nasal /n/ or sibilant /s/, in anticipation of the final rounded vowel.  This anticipatory coarticulation is across a syllable boundary.

  1. We see coarticulation across word boundaries so that we can conclude that words are not the basic building blocks of spoken language.

Consider /lEsru/ (“less true) for example, also has lip rounding initiated during the sibilant /s/, which is across a word boundary.  Therefore, we produce some words systematically differently, depending on the context of what comes next.  Words, therefore, cannot be invariant building blocks of spoken language.

 So far, we have considered anticipatory coariculation only in the context of anticipating rounded vowels.  In this situation, consonants that are neutral for lip rounding take on a rounded posture in anticipation of a rounded vowel that will occur later in the sequence.  Anticipatory lip rounding can be found in other contexts; for example, nasalization of vowels in anticipation of nasal consonants.

4.      Consider the contrasting pairs, /kQt/ “cat” and /kQn/ “can.”  The vowel, /Q/ is neutral for nasality because there are no nasal vowels in English.  That is, in English, it makes no difference whether a vowel is nasal or non-nasal because nasality does not make a phonemic difference in English. Vowels are never contrasted by the nasality feature in English.  Since it makes no perceptual difference if a vowel is nasalized, we typically will nasalize a vowel that is followed by a nasal consonant.  So, we’d expect the that the vowel in /kQn/ “can”  would be nasalized.  In this situation, the nasality feature of the final consonant is anticipated, and achieved as part of the vowel posture, giving the vowel a nasal quality which we all safely ignore.

Summary

Anticipatory Coarticulation:

 As a general rule, we will achieve the articulatory manifestation of any feature of a speech sound as early in the temporal sequence as the rules of phonology will allow, often well before we execute other articulatory features of a phoneme.

Please pay special attention to the features that make up speech sounds.  For all classifications of speech sounds, the features are much more important than you might imagine and you surely will need them in your work as speech-language pathologists and audiologists.

[ Back ]