supplementaryreadingunit2

supplementaryreadingunit2
Unit two
Text One
Direction: The following text is about how to identify phonemes. Please read aloud the tongue-twisters in the text, and try to see how fast you can read them. (M. Yaguello. Language Through the Looking Glass. Oxford: Oxford University Press, 1998:52-54)
Every human being is born with the same sound-producing (articulatory) ability. Everyone can utter a very wide range of sounds, as its clearly shown by the ―babbling‖of very young children. A baby can produce sounds which seem incredibly strange and outlandish to the people around it. Jakobson (1941) writes:
In its babble, a child can articulate an overall number of sounds which never come together in
a single language, not even in a whole family of language: consonants with vastly different points
of articulation, palatal and rounded, sibilant and fricative; then there are clicks, complex vowels,
diphthongs, etc.
For a child, these sounds are above all expressive, playful. Plunged into a world of meaningless sounds, the child plays with its voice as though the latter were a toy or a musical instrument. The sounds are also phatic –the babble informs the child, as it were, of its presence. At this stage, the sounds do not have communicative value: The child cannot yet organize them into message. But then, as it learns its mother tongue—as it listens to the people around it—the baby gradually identifies some order
within this sonic chaos, a system which effectively privileges certain sounds: the distinctive sounds (which convey but do not contain meaning) of the language: a language which the child, as yet, does not even realize it is learning.
Strangely enough, the child will progressively lose the ability to articulate the sounds which do not belong in this system, losing a natural ability in favor of a cultural one. What is more surprising still is that the child, in the early stage of language acquisition, loses not only those sounds which are foreign to the language it is acquiring, but also a good number of those which belong to the system. The child recognizes the, but has to re-learn how to articulate them . Learning this distinctive value of the phonemes thus entails a temporary regression. Accordingly, from age 2 to 4, there is a gap between the child’s passive competence (its ability to recognize the distinctive sounds of its language) and its active competence (its ability to reproduce them). If your child, for example, calls a sheep a ―seep‖, it does not mean that the child cannot hear the difference and the worst thing one can do is to start imitating the child and indulging in ―baby talk‖. When the world of sound switches from being phonetic to being phonological, the child has to learn how to identify the often subtle boundaries between sounds inasmuch as they generate differences in meaning. The more two phonemes resemble each other, the more the child will find it difficult to keep them apart; to distinguish, for example, between truck and chuck, or sing and thing: a fact which explains why we resort, partly in a spirit of play, partly for pedagogic purposes, to the classic children’s tongue-twisters.
She was a thistle sifter and sifted thistles through a thistle sieve
The sixth sheikh’s sixth sheep’s sick
I can think of six thin things and of six thick things too
Peter Piper picked a peck of pickled peppers
Around the rugged rocks the ragged rascal ran
Or the following ditty:
She sells sea-shells on the sea-shore.
The shells she sells are sea-shells I am sure,
For if she sells sea-shells on the sea-shore
Then I am sure she sells sea-shore shells.
These examples bear out the fact that the opposition between neighboring phonemes[?] as in shell and [s] as in sell or [θ]as in thin and [s] as in sin, is acquired with difficulty for young children (they usually pronounce tchuck or twuck for truck) and also for foreigners whose language contains a different kind of [r](the French) or no [r] at all.
The young child is extremely sensitive to minimal pairs and loves playing with paronymic sequences which bring together and juxtapose a group of very similar words. This accounts for the success, at nursery school, of ditties such as:
A flea and a fly in a flue
Were imprisoned, as what could they do?
Said the flea, ―let us fly‖
Said the fly, ―let us flee‖
So, they flew through a flaw in the flue.
Or:
Did you eever iver ever in your leaf life loaf
See the deevil divil devil kiss his weef wife woaf?
No, I neever niver never in my leaf life loaf
Saw the deevil divil devil kiss his weef wife woaf
While still young a child can acquire several foreign
languages, in addition to its mother tongue, and is able to operate several different, co-existing phonological systems. As time passes, however, the child finds it increasingly difficult to hear and therefore to articulate the sounds which do not have a distinctive value in its own language.
Text Two
Direction: The following text is about computers and phonology. Do you think that speech synthesis is possible? Why or why not? (E. Finegan. Language: Its Structure and Use (3rd ed.). New York: Harcourt Brace College Publishers, 1999:131-132) Several decades ago researchers thought it would be a matter of only a few years before computers would be able to recognize speech and to synthesize it. (You can think of this simply as being able to turn spoken language into print and print into spoken language.) Although there has been some progress on both fronts, the process has taken longer than most researchers anticipated. The reasons do not lie in any absence of sophistication in computers but rather in the complexity of the phonological processes that characterize human languages and in our inability to model in a computer just what speakers do when they produce spoken utterances and understand the utterances of others. For example, as we saw earlier in this chapter, natural speech occurs in a continuous stream and is not readily segmented without knowledge of the particular language involved. Just how human beings segment a continuous stream of spoken language into distinct words and recognize the sound segments in those words remains unclear.
The synthesis of speech by machine has also proved challenging. To understand why, focus on a string of sounds such as would occur in a simple word like sand. It would seem to be a
straightforward matter to put together a machine-generated form of [s?nd]:just get the machine to produce first a voiceless alveolar fricative, then the vowel [?], then the alveolar stop [d]. It seems
simple enough, but notice that when pronounce sand, its vowel quality differs markedly from the ―same‖ vowel in a word like hat. If a speech synthesizer produced the vowel of hat in the word sand, it would sound highly artificial. Likewise, if it produced the vowel of sand in the word hat, that too would seem very unnatural. You already know that the vowel of sand gets nasalized before the nasal stop that follows it. What happens in articulatory terms is that in anticipation of the following nasal consonant and as the vocal tract starts to move toward that nasal consonant, the vowel itself takes on nasal characteristics. Therein lies one challenge for speech synthesis—how to blend sounds into one another in the way that people do. Just as there is no separation between words in ordinary human speech, so there is no separation between sounds.
But the situation is even more complex. We have seen that a sound is essentially a bundle of phonetic features. Thus, we could think of the phonological form of sand as being not just the four segments [s?nd] but as the (partial list of) features given below to each segment: [s] [?] [n] [d]
voiceless voiced voiced voiced
alveolar low-front alveolar alveolar
fricative unrounded nasal stop
The phonetic characteristics of the segments of sand are more complicated than we have indicated, but the representation above will serve for our purposes. Consider that the articulation of the phonetic features in each segment does not start and end
at the same time as the others. In other words, the voicelessness of [s] doesn’t abruptly end and the voicing of [?] starts at exactly the same time as the fricative character of the consonant stops and the vowel character of [?] begins. The mouth and the other features of the vocal tract move continuously in the production of even a simple word like sand (as you can feel by saying the word and concentrating on your tongue movement).
If the aim of speech synthesis is to make artificial speech sound as natural as possible, a good deal more about the nature of phonetic realizations of underlying phonological forms will have been achieved.。