Track your comments!
[x]


When you register, comments on your articles and replies to your comments appear here. Register Now!

Sign in to your account
[x]

Not a Scientific Blogging member yet?

Register Now for a Free Scientificblogging.com Account

  • Customize your profile with pictures, banner, a blogroll and more.
  • Leave comments on articles, add other members to your friend lists, chat with people on the site.
  • Write blog posts that can be seen by hundreds of thousands of readers.

It's free and it only takes a minute!

Already a Scientific Blogging member?

Sign In Now

Banner
By Patrick Lockerby | June 19th 2009 11:37 AM | 4 comments | Print | E-mail | Track Comments
.

More The Chatter Box articles

All

About Patrick Lockerby

Retired engineer, 60+ years young.
Computer builder and programmer.
Linguist specialising in language acquisition and computational linguistics.
Interested in every human endeavour except the... Full Bio

A Science Of Human Language - Part #2


Quistic Grammar : A New Universal Grammar

In
Part #1 of this series, I suggested that a grammar heavily based in syntax was not sufficiently scientific as a general theory of how language functions.  In developing the current theory I shall try to demonstrate that various observations about human language can be tied together into an inclusive theory of how language functions.  The first, and to my mind most important observation about human language is its redundancy, its apparent inefficiency in the use of the resources of sounds and symbols.


The Function Of Redundancy In Human Language

An examination of samples of speech and writing in any language reveal that the core ideas being transmitted can often be coded in multiple ways.   A few examples from English should serve to demonstrate this.  Let us assume a scenario in which there are two conversers, a worker and an observer.  There is a single nail which has not yet been used and a single hammer.  There are many ways in which an intention to act in the future may be communicated by the worker to the other.

"I shall hit that nail with this hammer."
"I'm going to hit that nail with my hammer."
"I'm going to hit that nail with the hammer."
"I be hitting that nail with hammer."
"Me hit hammer nail."
"Hit hammer nail."
"Hammer nail"

In all of these examples, the logical components of the overall idea of a future event remain fairly constant.  A question naturally arises as to why language permits so much variation of style.  In the given context there is little room for ambiguity and, I suggest, little need for rhetorical flourishes.  Apart from the key words 'hammer', 'nail' and, possibly, 'hit' the other words appear to be a mere decoration.  They appear to the casual observer to be non-essential to the core logical idea.  This is the concept of redundancy as I shall use it within the scope of this theory.

Redundancy: any component of spoken or written speech is redundant if, on casual inspection, it is not essential to the meaning being conveyed.

Redundancy is only a surface appearance.  Components which do not directly convey or contribute towards the core meaning of a communication actually perform a vital error-handling role.


Error Handling In Human Language

The redundancy in speech extends far below the word and phoneme level.  The band of frequencies within which the human ear can detect speech sounds is far wider than needed in an ideal situation.  The range of intensities which can be discriminated is also greater than needed ideally.   This redundancy serves a useful purpose.   The frequency range can be severely reduced.  Speech can be volume-clipped, frequency-shifted, time-sliced with silent 'drop-outs' and mixed with out-of phase echoes.  And yet the original speech can still be  understood.

In the domain of language as words, rather than sounds, we can cope with variations from our personal view of 'ideal' use of language.  In both speech and writing we frequently overlook variations of usage and errors of many kinds: the brain performs an error-correction function which is, for the most part, entirely unconscious.

The automatic error-correction process operates by utilizing the redundancy in language: a component of meaning which fails to arrive via one method arrives via another method within the same language segment.  Language has apparently evolved to reduce the number of instances in which a whole message must be re-transmitted.


Pre-emptive Error-Handling

I suggest that the brain's error-handler performs another vital function: pre-emptive error-handling.  The surplus information in the apparently redundant components of language allow the brain to look ahead and determine the word category which will most plausibly be added next to the chain of words.  A simple example borrows a convention of cryptography: the three characters Alice, Bob and Carol.  Given a situational context of these three characters in a story, a simple form of error-handling is possible.

"I think you are entirely wrong.", said ___.


Within the context of the story there are only three plausible options for the missing word.  It may well be that the previous dialogue narrows the plausible choice down to only one name.   If just one letter is given, the choice is narrowed:  a or c could apply to both Alice and Carol, whilst the other available letters reduce the choice to one person.

In a narrative, a picture is built up of the characters' behaviours.  A narrative usually makes clear the characters' genders.  This background information may also be brought to bear in error-handling.  Let us suppose that Bob is male, and the others female.   It is now possible to demonstrate the handling of an error using redundancy with contextual information.

"I think you are entirely wrong.", said ____ as he left the room.

We already know that Bob is male, so in the absence of error the word 'he' is entirely redundant.  But in the situation where an error arises, the knowledge that Bob is male and that the missing name is male lets us fill in the blank with little or no mental effort.

In most cases of an insignificant speech or writing error the brain's error-handler is so efficient that we are unaware of seeing or hearing the error.   In the case of the look-ahead function, however, it is a common experience to 'pull out a word' ahead of a speaker.   In the case of quoted speech following in a sequential dialogue, we may well be mentally reaching for 'Bob' even before we become aware that the word is masked or missing.


Ambiguity Error-Handling

Ambiguity in language is common, and is a common cause of misunderstanding.   Ambiguity can arise where two words with different meanings sound the same or are spelled the same.   It can also arise where a particular construction can convey two - perhaps more - meanings:
"Police officer arrests drunken party goer in his underwear." 
It is common for ambiguity to be discovered by careful analysis of samples of language.  

Unless ambiguity is particularly striking, it is rarely noticed by the average language user.  Redundancy in language provides one of the error-correction mechanisms by which ambiguity may be reduced or eliminated.  In a story about a vehicle spares shop, the line: "She stood on the brakes to reach the headlights." may be unremarkable.   But if the sentence is to be taken out of context, the judicious addition of redundancy provides cues, i.e. aids in resolving potential ambiguities: "She stood on the boxes of brake pads in order to reach the headlights on the top shelf."


Words As Cues

It has long been observed that words seem to fall into two classes: the 'content words',  and the 'grammar words' of language.  It is often stated that the 'grammar words' are not carriers of meaning.  I suggest that they do indeed carry meaning.  The meaning is best expressed as a clue or cue in the problem posed to every reader or hearer of language: the problem of determining the intent of the speaker or writer.

I suggest that all words can convey meaning, and all words can function as cues.  Words thus fall naturally into two classes: those whose primary function is to transmit meaning, and those whose primary function is to assist in error-handling.  By the  insertion of cues into strings of words in accordance with the rules of conformity of a shared language, a 'speaking brain' makes the task of analysis and  evaluation much easier for a 'listening brain'.  Reduction of work load has obvious evolutionary survival value.

Part #3 introduces the concept of 'rules of grammar' as social 'rules of conformity'.

If you enjoyed this article, you may enjoy other articles about language in my blog, The Chatter Box.

Comments

Gerhard Adam's picture
"I think you are entirely wrong.", said ____ as he left the room.

In this example, we return to a problem we've discussed before about how this is a written language issue and would never occur (directly) in conversation.  

However, variations of this type of error do occur occasionally, when someone may not be able to think of a particular word and the listener provides it, seem to be very similar error handling mechanisms.  It is also clear from the conversational aspect of error handling that the listener is paralleling the conversation in their own mind, as well as constructing the images (or environment) the speaker is attempting to generate.

This fits with another discussion we've had regarding the role of abstraction because the vast majority of conversations necessarily involve experiential elements which don't exist in the environment in which the conversation is occurring.  In other words (see, I'm retransmitting the message), any reference to past or future is by definition an abstraction.  In most cases, a listener would consider it beyond obvious if someone were to narrate an activity that was being witnessed by the listener.  It seems that most conversations rarely deal with topics or issues which both parties are currently experiencing (except possible as a reminder of some shared past or projected future experience ... "remember this" ... "wouldn't it be fun to" ...etc. 

The most notable exception is when teaching is occurring.  In that instance it seems that the speaker is actively conversing in the present tense (i.e. "do this", or "solve x").  That may be more of a stretch.

Anyway .... as you can see, I'm one of those people where the written word (i.e. what I read) generates all kinds of additional thoughts that I'm perpetually compelled to write down and feed back.

logicman's picture
"I think you are entirely wrong.", said ____ as he left the room.

In this example, we return to a problem we've discussed before about
how this is a written language issue and would never occur (directly)
in conversation.

Once again, you've given me a lot of useful feedback, for which I thank you, Gerhard.

Imagine two people conversing in a noisy environment.  The occasional word will be partially or even completely masked by a louder than average noise.  It is in this situation that a more restrictive application of the 'rules' of language pays dividends.  Consider military communications with their use of restricted vocabularies.  This improves the correct-guess rate for words imperfectly heard.

It is also clear from the conversational aspect of error handling that
the listener is paralleling the conversation in their own mind, as well
as constructing the images (or environment) the speaker is attempting
to generate.

This is what I call an 'as if' scenario:  the listener is building a string of words 'as if' they were originating the string rather than following it.  I think it is more likely that the brain would use one means to serve two purposes rather than have a means to construct a string of output speech and a separate means to understand another's input speech.

Gerhard Adam's picture
Consider military communications with their use of restricted vocabularies.

Very true.   Alpha, Bravo, etc.

As a pilot you use the word "niner" instead of "nine" to avoid potential confusion with German.

I think the context and the receivers knowledge in the subject is important for what is redundant ant not
for ex. the phrase "Me hit hammer nail." can mean me hit hammer with nail if you don't now how hammers are used. And if you say something to a person and he doesn’t understand it can be because he doesn’t have the background knowledge so the sentence really get ambiguous so you have to say it again with more words. I am going to something like relative Redundancy and ambiguity we also make some likelihood thingy and dismiss al the unlikely meanings of the sentence

Add a comment

The content of this field is kept private and will not be shown publicly.
  • Allowed HTML tags: <sup> <sub> <a> <em> <strong> <center> <cite> <code> <TH><ul> <ol> <li> <dl> <dt> <dd> <img> <br> <p> <blockquote> <strike> <object> <param> <embed> <del> <pre> <b> <i> <table> <tbody> <div> <tr> <td> <h1> <h2> <h3> <h4> <h5> <h6> <hr> <iframe>
  • Lines and paragraphs break automatically.
  • Web page addresses and e-mail addresses turn into links automatically.
CAPTCHA
If you register, you will never be bothered to prove you are human again. And you get a real editor toolbar to use instead of this HTML thing that wards off spam bots.