Re: Fw: Intro to natural language processing
Rob,
The crucial point is the definition of "symbol",
which Deacon adopted from Peirce. Following is a
brief definition from one of Peirce's manuscripts:
1. There are three kinds of signs. Firstly, there are
likenesses, or icons, which serve to convey ideas of
the things they represent simply by imitating them.
2. Secondly, there are indications, or indices, which
show something about things, on account of their being
physically connected with them. Such is a guidepost,
which points down the road to be taken, or a relative
pronoun, which is placed just after the name of the
thing intended to be denoted, or a vocative exclamation,
as "Hi! there," which acts upon the nerves of the person
addressed and forces his attention.
3. Thirdly, there are symbols, or general signs, which have
become associated with their meanings by usage. Such are
most words, and phrases, and speeches, and books, and
libraries.
For the full article from which this excerpt was taken, see
http://www.iupui.edu/~peirce/ep/ep2/ep2book/ch02/ep2ch2.htm
What is a Sign?
A key difference between index and symbol is that an index
commonly occurs in the presence of its referent. To use
Peirce's examples, the guidepost and the exclamation both draw
attention to the thing they are indices of -- the road or the
person who called -- both of which were present. Symbols,
however, are most commonly used when their referent is absent.
In effect, the symbol is a virtual placeholder for the absent
entity, and it enables the possibility of talking about events
and situations that are remote in time and place. You can
point (index) to things in the here and now (indexicals), but
you can't make plans or promises (future) without symbols.
Deacon's claim, which he develops in Chapter 3 in about 30 pages,
is that higher mammals and birds can easily learn icons and
indices, but not symbols. And by symbol, he is using Peirce's
definition. He discusses the ability of dogs to learn large
numbers of verbal commands, but he shows that all those signs
can be interpreted as indexical (in Peirce's sense) and none
as symbolic.
Deacon also cites evidence of work with chimps and bonobos,
most of whom learn their sign patterns in the same indexical
style as dogs (although their greater brain power enables them
to learn larger numbers of signs). However, he also cites
evidence that indicates the possibility, although rare and
difficult to achieve without extensive training, of some few
chimps and bonobos who crossed the threshold from index to symbol.
Those very few who have crossed that threshold are able to learn
vastly more combinations with far greater speed and accuracy
than their comrades.
The key to the remainder of the book is Peirce's definition
of "symbol" together with Deacon's review of the evidence
of chimps and bonobos learning signs and some very few
learning symbols. Once symbols are acquired, rudimentary
language is possible, and more complex langage with complex
grammar can evolve.
John