As a kind of human behavior, language covers several different substances (materials) which humans have contact with. Language can be written, spoken, or signed, and these channels are manipulated through bodily actions. Language form divides and relates these substances to each other and to their meanings. And meanings are yet another substance. But language substance is often gooey -- its format is often not obvious, not even to its users.
Articulatory Gestures
The signs of speech are not sounds, any more than the signs of sign language are light. Each morpheme (word) is an articulatory gesture, oral movement representing meaning. These are the signs of speech. Unlike sign language gestures, the signs (gestures) of speech are analyzable into subgestures (letters) which are detectable as sounds. But vocal chord vibrations make up a very small part of the speech signal, most of the distinct sounds being modifications of those vibrations caused by the shape the vocal tract. So, in real-time language, what we perceive are gestures, whether they are made with the mouth or with the hands, as in sign language. Sound and light are only the signals by which we perceive the gestures. For more details, click here.Writing
Ever since writing was invented, it has had a powerful influence on how we think about language. When people say English is a difficult language they usually mean English spelling is difficult or It's hard to write a good paragraph. They mean that English is hard to write, not hard to speak. Strange to think, then, that writing has only been around for about 5,000 years, but people have spoken for at least ten to twenty times longer. Writing influences the way we speak -- have you heard people pronounce the t in often? You won't find that "t" in the pronunciation of often in the dictionary. Writing even influences some linguists' otherwise objective analyses. Why?Writing is magic. Imagine how ancient people must have felt when someone miraculously spoke someone else's words weeks after they had disappeared into the air! Or when they read a message from someone perhaps miles away. Imagine that a small elite group of people were the only ones who could read the wisdom of the ancients -- what authority they must have had! Now try to imagine modern society without writing. Mathematics and science would be virtually impossible. And without them there wouldn't have been the digital technology which appear to be making reading obsolete. How could we even produce radio or television programs without scripts? Writing has truly transformed society.
But how was writing invented? Writing started as pictures, icons drawn on stones to record historic events or financial transactions. Slowly the pictures, which perhaps had stood for an entire story, began to represent words. This was accomplished by introducing the idea of a phonetic. Sometimes an icon would be used as a phonetic symbol. If a word like be was too abstract to be a picture, an icon of a word with a similar sound could be substituted, like the picture of a bee. In addition, a signific (a hint of the meaning), could be combined with the phonetic, making it absolutely clear what word was meant. The picture of a bee combined with an equal sign "=" just about has to mean be. This kind of writing, still existing in Chinese, was independently invented only two or three times so far as linguists know.
When speakers of other languages picked up the idea of writing, the combinations of phonetics and signfics did not make any sense, since their languages were different. So sometimes, a new language used a limited set of the phonetics to stand strictly for sounds. Some languages took symbols for whole syllables, as in Japanese hiragana/katakana writing. Others, like Greek, began to use letters.
Letters are the usually the smallest symbols of phonetic writing, and belong to an alphabet. They often correspond more-or-less to articulatory subgestures, the tiny tasks we perform when we articulate (speak) words. Some common alphabets in use today are: Roman, Greek, Hebrew, Arabic, Cyrillic, Korean Hangul, Devanagari.
Sign Language
Sign language is a system of gestures of the hands, face and upper body, used mostly by deaf people instead of spoken language. American Sign Language (ASL) is one such system. The symbols (signs) of ASL are not the words or meanings of English. They are as far from English as those of French, maybe farther. The order of ASL signs is often different from English word order as well.Most ASL signs consist of one or two gestures, but not longish strings of subgestures as in an English word. Some signs may remind you of the things they stand for, but you could never guess what they mean. If you're a non-signer and you've seen people signing in ASL, you know.
A very few signs are obvious: the sign for banana is made by miming the peeling of your index finger like a banana. To the extent that ASL has these obvious signs, it is partly understandable at least by signers of other sign languages.
According to our definition, sign language is a language. It has symbols with general meanings, the symbols are transmitted (by light waves), and they can be used to communicate or store information, and to think. It can translate the Bible. But sign language is less arbitrary than spoken language, since many of its signs at least resemble the ideas they stand for. Certainly its version of the double articulation is different from that of spoken language. And for those linguists who insist that a language must be vocal, Sign Language is by their definition not a language. American Sign Language possesses a manual alphabet, a separate set of hand gestures which can be used to spell words in English, usually proper names.
Besides ASL, there are many other sign lanauges used by deaf people in other counties. Also, non-deaf native peoples of North America and Australia have used sign languages in addition to their spoken languages.
As the organ of intelligence and thinking, the brain is essential to language. As the midpoint between speech perception and speech production, it must relate and separate our symbolic and non-symbolic perceptions. It determines when we speak and when we pay attention to speech. And as the endpoint of the speech chain, it does a great deal of speech processing in both perception and production.
Psycholinguists have known for years that certain speech functions are located in specific areas of the brain. This was determined by comparing the various aphasias (speech abnormalities) with various damaged areas in the brain. Frustratingly, such research has been limited to what can be discovered in the cadavers of people who were known to have had aphasias. More recently, MRI imaging and other interactive methods are now providing more and more information on what happens where in the brain and how.
The major location of speech production is Broca's area. Found in the back of the brain's left frontal lobe, if this area is damaged, speech is no longer fluent, and the most frequent words are often omitted. The major location of speech perception is Wernicke's area, found in the back of the temporal area. When this area is damaged, speech comprehension suffers. Not surprisingly, Broca's area is near the area that controls the muscles producing articulatory gestures, and Wernicke's area is near the area that receives auditory and visual input. Although somewhat distant from each other, Broca's and Wernicke's areas are connected by the arcuate fasciculus, perhaps involved in monitoring the production of one's own speech.
Currently, a major area of scientific debate centers on just how modular speech is: how is the processing of various aspects of language form specialized into separate regions of the brain, if at all? Another age-old debate concerns how much language form can influence our thinking and our perception of the universe. The viewpoint that thought is influenced by language is known as the Sapir-Whorf hypothesis.
Speech perception is very different from reading. In a printed text, the letters are neatly arranged next to -- but not touching -- each other. Words are separated from each other by spaces. And the reader sets his or her own pace for input. None of this is true for speech. In spoken language, the sounds blend into one another as the mouth adjusts itself from the finish of one articulatory subgesture (letter) to the start of the next. Exactly the same transitions occur between words -- needless to say, we do not pause or pronounce a "space" between words in a phrase. The same is true of sign language.
Speech sounds are produced so quickly that if they were counted as "beats" they would sometimes produce a vibration audible as a very low note. This is one reason that some psycholinguists believe that speech perception is different from ordinary sound perception, almost like a sixth sense. Another reason is our categorical perception of speech sounds. It is possible, for instance, to electronically synthesize a sound midway between L and R. Speakers of English or other languages that distinguish these sounds will never identify this as an intermediate sound, but will always hear either L or R. Japanese speakers, who don't distinguish the two sounds in their language, will hear their own combination L/R sound. On the other hand, Japanese speakers will hear this same sound when they hear the pure L or pure R.
Whereas reading takes place only through the visual sense, speech is perceived through multiple senses, not just the auditory channel. Everyone is able to lip-read well enough to greatly improve understanding by looking at a speaker's mouth at a noisy party. People also perceive and monitor their own speech production through the tactile (touch) and kinesthetic (motor) senses. In addition to helping us perceive speech, these multiple senses may also help us segment speech into its double articulation format.
Although speech can certainly be analyzed into its component sounds or subgestures, and then into words or morphemes, it is questionable whether speech is perceived or even produced as such. For instance, the sequence of English words of the is heard twice as often as the word house. In fact, the French version de le became the single word du hundreds of years ago. Contrary to some linguistic theories, it is doubtful that any French speaker or listener mentally transforms du to de le or vice versa. More likely, hearing de le instead of du would simply confuse a native French speaker. Rather than building up language from its smallest parts, children probably learn languages by hearing strings of sounds in appropriate contexts, imitating them, and then learning where they can be modified by comparing them to other strings they have heard. Analysis, rather than synthesis. For instance, if a child has already heard I'mgoing, you'regoing and I'mhungry, the child will probably guess that it makes sense to say you'rehungry. Later on, children learn what words rhyme and eventually learn to read, using an alphabet. However, it is usually a struggle for children to sound out words phonetically and learn to spell and recognize new words, even in a language with few spelling exceptions, such as Spanish. If children learn language in an unanalyzed format, it seems likely that, as adults, they continue to speak and perceive language at least partly in an already-synthesized format.
Speech is probably not synthesized from its smallest elements in production any more than it is in perception. Especially when uttering common phrases like how are you!, you are not likely to re-construct them from the separate words how + are + you. That is one reason why such expressions become abbreviated to hi ya! and then simply to hi! But even more complicated sentences like I really think you're wrong may actually be composed of I really think _ wrong + you're. Instead of you're, the word he's could be substituted in the blank. Many linguistic theories break the sentence first into: I + really think you're wrong, then the rest into really + think you're wrong, then into think + you're wrong, etc., etc. If it were produced from this many pieces with this kind of synthesis, human speech would be very slow indeed!
It seems more likely that speech is composed of frames like the above with various words or expressions substituted in the blanks. One bit of evidence for this is the fact that people pause longer before less frequent words. In the phrase of the intervention you will not hear people pause between of and the, but before intervention, as they try to recall the long, infrequent word. This is so, despite the claim of many grammatical theories that there is a more significant break between of and the than between the and intervention.
In any case, speech production probably takes place on many levels simultaneously. In Spanish, speakers have to determine the "gender" of a noun before they can use the correct definite article (translation of the English word the), either el or la. Now, when they pause to recall a noun, Spanish speakers pause by prolonging el… or la…, rather than using a word like umm… as in English. But this seems like a strange gamble, because they ought to be as likely as not to come up with a noun of the wrong gender and have to take back the article: del… de la intervención. However, in practice, they use the right gender four out of five times when they hesitate on the article while recalling a noun. The author has verified this statistically on a transcription of about a million words of Spanish conversations. Whatever the cause of the hesitation, this seems to show that the genders of Spanish nouns are at least sometimes recalled separately from their pronunciations, despite the fact that the gender does not have any real-world meaning.
Many factors influence speech production,
but the logician H.P. Grice has come of with several "conversational maxims",
or guidelines, which seem to universally represent good practice
in speech. They are: QUANTITY -- don't use any more words than necessary,
QUALITY -- speak the truth, RELATION -- keep your conversation relevant
to the topic at hand, MANNER -- be clear and to the point.
Although people don't always follow these guidelines, especially in spoken
language, it seems we can all agree that it's frustrating when they don't.
The maxim of QUANTITY seems to relate to information
theory, but, so far, the others have not been easily expressed mathematically,
or been detectable automatically by computer.
Sound is the most important signal by which speech is perceived, although it's not the only channel. Sound is actually only a rapid vibration of the air. The faster the vibration, the higher the frequency. The stronger the vibration, the louder the sound. Frequency and loudness through time are all that needs to be recorded to reproduce a sound sequence on a cassette tape or CD.
Linguists sometimes record the frequency
and loudness of speech on a special device called a spectrograph.
The resulting spectrograms are printed out and the result is a kind of
visible speech. Visible
Speech is also the name of the book from which the illustrations
of this section have been copied with the kind permission of Dover Publications.
The first spectrogram shows the
following sentence fragment: Any bright young man who still thinks….
Reading it from left to right, you can see that speech sounds are not equally
spaced like normal written letters. Reading up and down, you
can see where various frequencies have any loudness. That is,
the multiple horizontal dark bands show that speech sounds consist of multiple
simultaneous frequencies. The higher the frequency, the higher up
on the spectrogram the dark band will appear. The darker the band,
the louder the sound at that frequency. You may not be aware that
you hear several frequencies simultaneously in speech. However,
they are the result of the reinforcement that the various shapes of the
mouth give to the more-or-less pure tone of the vocal cords.
In other words, you unconsciously detect most articulatory
gestures by the sound reflection on the inside of another speaker's
mouth, much in the same way you can tell about how much water is in your
teacup by hearing the rising tone of the pouring sound produced as the
cup fills up.
Another striking thing about the spectrogram is how often the horizontal bands are interrupted. These interruptions are usually consonants. For example, the stop consonants B and T in bright stop up the mouth temporarily with the lips and the tongue respectively. They cut off the sound, and leave only a very light color or just one band on the bottom of the spectrogram, as with B. Fricatives, the hissing consonants, leave a fuzzy haze, as in the S at the very end of our spectrogram, on the end the word thinks. This hissing is sometimes called white noise.
The third surprising aspect of spectrogram
is how one sound blends into another. Because we are used to printing,
we expect something to clearly separate the sounds as if they were letters.
Speech is more like handwriting than printing. But in speech, one
sound not only blends into the next, but usually the sounds themselves
are adjusted by their neighbors. This can be seen in the second
illustration, a set of spectrograms representing the phrase we owe
you.
The upper spectrogram (a) is taken
from a normal pronunciation of the phrase, the lower spectrogram (b) represents
the same phrase spoken by carefully separating each word: we, owe,
you. Compare the areas that represent the word
owe
in each spectrogram. You can see a smile-shaped upper band
on the first one, but on the second spectrogram, this upper band is flat
(and partly blended into the lower band). This is because the
transitions to the neighboring words are not pronounced in the second one.
You can hear the transitions for yourself if you pronounce we-owe-you
slowly, but connectedly, without pausing or stopping your voice.
If you listen carefully, you can hear something like yo between
we
and owe, and oy between owe and
you.
The point of all this is that there is structure in language, but it is not obvious or easy to detect automatically. The double articulation saves us from having to physically create a new sound for every word, but that's all it does. It is not there to save brain storage in speech production or perception. Although a word or even a phrase is made up of articulatory subgestures, it may not be stored in the brain as such. Rather than an instruction to assemble a set of words and subgestures, a phrase may be stored as a whole, together with some information on where it can be divided if necessary.