Re: Learning a language
From: Nathan Sanders (nsanders.DIE.SPAM_at_wso.williams.edu)
Date: 07/07/04
- Next message: Joseph W. Murphy: "Re: The Phaistos Disk, side A, according to Fischer"
- Previous message: Qakare: "Re: The Phaistos Disk, side A, according to Fischer"
- In reply to: Patrick Powers: "Re: Learning a language"
- Next in thread: LEE Sau Dan: "Re: Learning a language"
- Messages sorted by: [ date ] [ thread ]
Date: Wed, 07 Jul 2004 17:13:11 GMT
In article <9511688f.0407070303.5f3a208d@posting.google.com>,
frisbieinstein@yahoo.com (Patrick Powers) wrote:
> If I may be so bold, I can suggest a measure of language complexity.
> You could measure how easy it is to teach a computer the language.
> Teaching people is trickier because so much depends on what they
> already know.
This is a reasonable start on trying to define linguistic complexity,
but I see some potential problems:
* How much of the language will be "taught" to the computer? Just
phonetics, phonology, morphology, and syntax? What about semantics
and pragmatics (which is where humans seem to do much better than
computers do)?
* What will be measured: How long it takes to create the code for the
language? How many bytes the finished code takes up? How many
seconds it takes the computer to compile the interpretation program
for the langauge? How long it takes a learning algorithm to reproduce
your code after being exposed to linguistic data? Something else?
* Once you find a way to measure the complexity of a language for
computers, how can we be sure that this measurement will carry over to
humans? What's easy/complex for a computer may not be easy/complex
for a human: multiplying large numbers and memorizing long strings
(easy for computers, hard for humans) versus recognizing faces and
finding short paths in the traveling salesman problem (hard for
computers, easy for humans).
Some linguists are in fact already doing essentially what you're
suggesting. However, they are focusing on very narrow aspects of
language (say, irregular verbs in English) and not on whole languages
(especially the entire syntactic, semantic, and pragmatic components,
where our knowledge is very limited, even for well-studied languages
like English, but especially for languages like Esperanto which have
no monolingual community).
Linguists simply don't understand all of the intricacies of language
well enough to quantify them in a way that allows them to be reliably
and consistently compared with each other as "simple" or "complex".
Any statement to the contrary is either a falsehood or an incredible
breakthrough that should be appearing in academic circles around the
world any day now.
If such a statement appears on Usenet, from admitted non-linguists no
doubt, I'm more likely to go with the former rather than the latter.
Nathan
-- Nathan Sanders Linguistics Program nsanders@wso.williams.edu Williams College http://wso.williams.edu/~nsanders Williamstown, MA 01267
- Next message: Joseph W. Murphy: "Re: The Phaistos Disk, side A, according to Fischer"
- Previous message: Qakare: "Re: The Phaistos Disk, side A, according to Fischer"
- In reply to: Patrick Powers: "Re: Learning a language"
- Next in thread: LEE Sau Dan: "Re: Learning a language"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
